Compare commits

...

13 Commits

Author SHA1 Message Date
KyuubiYoru cc5793f935 feat(release): add reproducible signed artifacts (#19)
quality-gate / quality (push) Failing after 1m50s
quality-gate / container (push) Has been skipped
2026-07-16 17:48:21 +02:00
KyuubiYoru 07004cd75f docs(evidence): record clean capacity candidate (#18)
quality-gate / quality (push) Failing after 1m28s
quality-gate / container (push) Has been skipped
2026-07-16 16:11:39 +02:00
KyuubiYoru cf14836d48 fix(operations): require clean candidate provenance (#18) 2026-07-16 16:04:25 +02:00
KyuubiYoru 609dad7cf1 feat(operations): add capacity and resilience gates (#18) 2026-07-16 15:57:01 +02:00
KyuubiYoru 08729ae25c feat(deployment): add secure Linux runtime (#17)
quality-gate / quality (push) Failing after 1m9s
quality-gate / container (push) Has been skipped
2026-07-16 15:03:04 +02:00
KyuubiYoru be732de7c9 feat(server): add observability and operator controls (#16)
quality-gate / quality (push) Failing after 1m1s
2026-07-16 13:22:16 +02:00
KyuubiYoru 88ef946af5 feat(server): harden hostile input and overload behavior (#15)
quality-gate / quality (push) Failing after 1m5s
2026-07-16 12:37:38 +02:00
KyuubiYoru 2ff7cd6d9d test(integration): add deterministic NAT topology harness (#14)
quality-gate / quality (push) Failing after 1m3s
2026-07-16 11:50:53 +02:00
KyuubiYoru 7e3be2cad1 feat(tooling): add standalone rendezvous test client (#25)
quality-gate / quality (push) Failing after 1m3s
2026-07-16 11:05:56 +02:00
KyuubiYoru 94aba8a3bb feat(client): standardize connection outcomes (#13)
quality-gate / quality (push) Successful in 59s
2026-07-16 10:18:41 +02:00
KyuubiYoru b4b6072fe1 feat(client): add rendezvous traversal coordinators (#12)
quality-gate / quality (push) Successful in 56s
2026-07-16 08:39:05 +02:00
KyuubiYoru 6d076c281a feat: implement authenticated NAT mediator (#11)
quality-gate / quality (push) Successful in 59s
Closes #11
2026-07-16 07:37:02 +02:00
KyuubiYoru 1baa1055dc feat: implement scoped join attempts and tickets (#10)
quality-gate / quality (push) Successful in 1m1s
Closes #10
2026-07-16 06:56:30 +02:00
179 changed files with 25658 additions and 579 deletions
+13
View File
@@ -0,0 +1,13 @@
.git
.gitea
.idea
.vs
.codex
.agents
**/bin
**/obj
TestResults
deploy/compose/secrets
deploy/compose/.smoke.env
docs
tests
+136 -2
View File
@@ -14,16 +14,34 @@ jobs:
timeout-minutes: 15 timeout-minutes: 15
steps: steps:
- name: Check out repository - name: Check out repository
uses: actions/checkout@v4 uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
with:
fetch-depth: 0
- name: Install .NET SDK - name: Install .NET SDK
uses: actions/setup-dotnet@v4 uses: actions/setup-dotnet@67a3573c9a986a3f9c594539f4ab511d57bb3ce9 # v4.3.1
with: with:
dotnet-version: 10.0.301 dotnet-version: 10.0.301
- name: Restore locked dependencies - name: Restore locked dependencies
run: dotnet restore Rendezvous.slnx --locked-mode run: dotnet restore Rendezvous.slnx --locked-mode
- name: Verify dependency licenses and reviewed transport pin
run: python3 eng/release_artifacts.py policy --root .
- name: Reject vulnerable direct or transitive packages
shell: bash
run: |
set -euo pipefail
dotnet package list --project Rendezvous.slnx \
--vulnerable --include-transitive --no-restore --format json \
>"${RUNNER_TEMP}/nuget-vulnerabilities.json"
python3 eng/release_artifacts.py audit \
--input "${RUNNER_TEMP}/nuget-vulnerabilities.json"
- name: Enforce compatibility version bumps
run: ./scripts/check-compatibility.sh origin/main
- name: Verify formatting and analyzers - name: Verify formatting and analyzers
run: dotnet format Rendezvous.slnx --verify-no-changes --no-restore run: dotnet format Rendezvous.slnx --verify-no-changes --no-restore
@@ -35,3 +53,119 @@ jobs:
- name: Test - name: Test
run: dotnet test Rendezvous.slnx --configuration Release --no-build run: dotnet test Rendezvous.slnx --configuration Release --no-build
- name: Run quick capacity and resilience gate
run: ./scripts/run-capacity-gate.sh
- name: Test privileged Linux namespace topology when available
shell: bash
run: |
set -euo pipefail
probe="rendezvous-probe-$$"
suffix="$(( $$ % 100000 ))"
bridge="rvb${suffix}"
veth_root="rvr${suffix}"
veth_peer="rvp${suffix}"
cleanup_probe() {
if [[ -n "$veth_root" ]]; then
ip link delete "$veth_root" >/dev/null 2>&1 || true
fi
if [[ -n "$bridge" ]]; then
ip link delete "$bridge" >/dev/null 2>&1 || true
fi
if [[ -n "$probe" ]]; then
ip netns delete "$probe" >/dev/null 2>&1 || true
fi
}
trap cleanup_probe EXIT
if command -v ip >/dev/null 2>&1 \
&& command -v iptables >/dev/null 2>&1 \
&& command -v sysctl >/dev/null 2>&1 \
&& ip netns add "$probe" 2>/dev/null \
&& ip link add "$bridge" type bridge \
&& ip link add "$veth_root" type veth peer name "$veth_peer" \
&& ip link set "$veth_root" master "$bridge" \
&& ip link set "$veth_peer" netns "$probe" \
&& ip netns exec "$probe" sysctl -q -w net.ipv4.ip_forward=1 \
&& ip netns exec "$probe" iptables -t nat -A POSTROUTING -o "$veth_peer" -j MASQUERADE \
&& ip netns exec "$probe" iptables -A FORWARD -i "$veth_peer" -o lo \
-m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT; then
ip link delete "$veth_root"
veth_root=""
ip link delete "$bridge"
bridge=""
ip netns delete "$probe"
probe=""
results="${RUNNER_TEMP:-/tmp}/rendezvous-netns-results"
mkdir -p "$results"
RENDEZVOUS_RUN_NETNS_TESTS=1 dotnet test Rendezvous.slnx \
--configuration Release \
--no-build \
--filter FullyQualifiedName~PrivilegedLinuxNatNamespacesCompleteDirectTrafficAcrossSeparateObservedEndpoints \
--logger "trx;LogFileName=netns.trx" \
--results-directory "$results"
grep -q 'testName="[^"]*\.PrivilegedLinuxNatNamespacesCompleteDirectTrafficAcrossSeparateObservedEndpoints"' \
"$results/netns.trx"
else
echo "Network namespaces/NAT tooling unavailable; deterministic loopback topology remains the required gate."
fi
container:
needs: quality
runs-on: ubuntu-latest
timeout-minutes: 15
steps:
- name: Check out repository
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- name: Install .NET SDK
uses: actions/setup-dotnet@67a3573c9a986a3f9c594539f4ab511d57bb3ce9 # v4.3.1
with:
dotnet-version: 10.0.301
- name: Build deployment diagnostic
run: |
dotnet restore src/FinalFactory.Rendezvous.TestClient/FinalFactory.Rendezvous.TestClient.csproj --locked-mode
dotnet build src/FinalFactory.Rendezvous.TestClient/FinalFactory.Rendezvous.TestClient.csproj --configuration Release --no-restore
- name: Build and exercise hardened container
shell: bash
run: |
set -euo pipefail
compose_file="deploy/compose/compose.yaml"
secret="deploy/compose/secrets/signing-key"
cleanup() {
RENDEZVOUS_UID=1654 RENDEZVOUS_GID=1654 \
docker compose -f "$compose_file" down --volumes >/dev/null 2>&1 || true
rm -f "$secret"
}
trap cleanup EXIT
install -d -m 0700 deploy/compose/secrets
openssl rand -out "$secret" 32
chmod 0444 "$secret"
export RENDEZVOUS_UID=1654
export RENDEZVOUS_GID=1654
export SOURCE_REVISION_ID="$GITHUB_SHA"
docker compose -f "$compose_file" build \
--build-arg SOURCE_REVISION_ID="$SOURCE_REVISION_ID"
docker compose -f "$compose_file" up --no-build --detach
container_id="$(docker compose -f "$compose_file" ps -q rendezvous)"
test -n "$container_id"
test "$(docker inspect --format '{{.Config.User}}' "$container_id")" = "1654:1654"
test "$(docker inspect --format '{{.HostConfig.ReadonlyRootfs}}' "$container_id")" = "true"
test "$(docker inspect --format '{{range .Mounts}}{{if eq .Destination \"/app/appsettings.Production.json\"}}{{.RW}}{{end}}{{end}}' "$container_id")" = "false"
test "$(docker inspect --format '{{range .Mounts}}{{if eq .Destination \"/run/secrets/rendezvous-signing-key\"}}{{.RW}}{{end}}{{end}}' "$container_id")" = "false"
for attempt in {1..100}; do
if curl --fail --silent http://127.0.0.1:8080/health/ready >/dev/null 2>&1; then
break
fi
if (( attempt == 100 )); then
docker compose -f "$compose_file" logs rendezvous
exit 1
fi
sleep 0.1
done
./scripts/smoke-deployment.sh
docker compose -f "$compose_file" stop --timeout 40 rendezvous
test "$(docker inspect --format '{{.State.Running}}' "$container_id")" = "false"
test "$(docker inspect --format '{{.State.ExitCode}}' "$container_id")" = "0"
+193
View File
@@ -0,0 +1,193 @@
name: immutable-release
on:
push:
tags:
- "v*.*.*"
concurrency:
group: release-${{ gitea.ref_name }}
cancel-in-progress: false
jobs:
release:
runs-on: ubuntu-latest
timeout-minutes: 45
environment: production
steps:
- name: Check out immutable tag
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
with:
fetch-depth: 0
- name: Install pinned .NET SDK
uses: actions/setup-dotnet@67a3573c9a986a3f9c594539f4ab511d57bb3ce9 # v4.3.1
with:
dotnet-version: 10.0.301
- name: Install pinned Buildx and BuildKit
uses: docker/setup-buildx-action@e468171a9de216ec08956ac3ada2f0791b6bd435 # v3.11.1
with:
version: v0.35.0
install: true
driver-opts: image=moby/buildkit:v0.25.2@sha256:0f63d66f8d2de0bd16438284831a3e9ee6ca7cd57b6eb3ed6e38a7a456590fa7
- name: Validate tag and produce reproducible artifacts
shell: bash
run: |
set -euo pipefail
version="${GITHUB_REF_NAME#v}"
./scripts/check-release-tag.sh "$GITHUB_REF_NAME"
previous_tag="$(git tag --merged HEAD^ --list 'v*.*.*' --sort=-version:refname | sed -n '1p')"
if [[ -n "$previous_tag" ]]; then
./scripts/check-compatibility.sh "$previous_tag"
elif [[ -n "$(git tag --list 'v*.*.*' | sed -n '1p')" ]]; then
echo "No prior release tag is an ancestor of $GITHUB_REF_NAME." >&2
exit 1
else
./scripts/check-compatibility.sh __initial_release_without_base__
fi
release_builder="rendezvous-release-builder:${GITHUB_SHA}"
docker buildx build \
--platform linux/amd64 \
--file eng/release-builder.Dockerfile \
--target release-builder \
--load \
--tag "$release_builder" .
mkdir -p "${RUNNER_TEMP}/release-home" "${RUNNER_TEMP}/nuget"
docker run --rm \
--user "$(id -u):$(id -g)" \
--env HOME="${RUNNER_TEMP}/release-home" \
--env NUGET_PACKAGES="${RUNNER_TEMP}/nuget" \
--volume "$GITHUB_WORKSPACE:/source" \
--volume "${RUNNER_TEMP}:${RUNNER_TEMP}" \
--workdir /source \
"$release_builder" \
./scripts/build-release.sh "$version" "${RUNNER_TEMP}/release/$version"
./scripts/verify-real-consumers.sh "$version" "${RUNNER_TEMP}/release/$version"
echo "RENDEZVOUS_VERSION=$version" >>"$GITHUB_ENV"
echo "RENDEZVOUS_RELEASE_DIR=${RUNNER_TEMP}/release/$version" >>"$GITHUB_ENV"
echo "RENDEZVOUS_RELEASE_BUILDER=$release_builder" >>"$GITHUB_ENV"
- name: Build exact container candidate
shell: bash
run: |
set -euo pipefail
export SOURCE_DATE_EPOCH="$(git show -s --format=%ct HEAD)"
common=(
--no-cache
--pull=false
--provenance=false
--platform linux/amd64
--build-arg SOURCE_DATE_EPOCH="$SOURCE_DATE_EPOCH"
--build-arg SOURCE_REVISION_ID="$GITHUB_SHA"
)
release_tag="git.finalfactory.de/heikyu/rendezvous:${RENDEZVOUS_VERSION}"
image_one="${RUNNER_TEMP}/rendezvous-image-1.tar"
image_two="${RUNNER_TEMP}/rendezvous-image-2.tar"
docker buildx build "${common[@]}" --tag "$release_tag" \
--output "type=docker,dest=$image_one,rewrite-timestamp=true" .
docker buildx build "${common[@]}" --tag "$release_tag" \
--output "type=docker,dest=$image_two,rewrite-timestamp=true" .
cmp --silent "$image_one" "$image_two"
docker load --input "$image_one"
candidate_id="$(docker image inspect --format '{{.Id}}' "$release_tag")"
buildkit_version="$(docker buildx inspect --bootstrap | sed -n 's/.*BuildKit version: *//p' | sed -n '1p')"
docker run --rm \
--user "$(id -u):$(id -g)" \
--volume "$GITHUB_WORKSPACE:/source" \
--volume "${RUNNER_TEMP}:${RUNNER_TEMP}" \
--workdir /source \
"$RENDEZVOUS_RELEASE_BUILDER" \
python3 eng/release_artifacts.py record-container-build \
--provenance "$RENDEZVOUS_RELEASE_DIR/release-provenance.json" \
--buildx-version "$(docker buildx version)" \
--buildkit-version "$buildkit_version" \
--image-id "$candidate_id"
- name: Stage HTTP registration, browse, and authenticated UDP traversal
shell: bash
run: |
set -euo pipefail
secret="deploy/compose/secrets/signing-key"
cleanup() {
RENDEZVOUS_UID=1654 RENDEZVOUS_GID=1654 RENDEZVOUS_IMAGE="git.finalfactory.de/heikyu/rendezvous:${RENDEZVOUS_VERSION}" \
docker compose -f deploy/compose/compose.yaml down --volumes >/dev/null 2>&1 || true
rm -f "$secret"
}
trap cleanup EXIT
install -d -m 0700 deploy/compose/secrets
openssl rand -out "$secret" 32
chmod 0444 "$secret"
export RENDEZVOUS_UID=1654
export RENDEZVOUS_GID=1654
export RENDEZVOUS_IMAGE="git.finalfactory.de/heikyu/rendezvous:${RENDEZVOUS_VERSION}"
docker compose -f deploy/compose/compose.yaml up --detach --no-build
for attempt in {1..100}; do
curl --fail --silent http://127.0.0.1:8080/health/ready >/dev/null 2>&1 && break
if (( attempt == 100 )); then
docker compose -f deploy/compose/compose.yaml logs rendezvous
exit 1
fi
sleep 0.1
done
./scripts/smoke-deployment.sh
- name: Scan candidate for high and critical vulnerabilities
uses: aquasecurity/trivy-action@57a97c7e7821a5776cebc9bb87c984fa69cba8f1 # v0.35.0, post-incident safe SHA
with:
image-ref: git.finalfactory.de/heikyu/rendezvous:${{ env.RENDEZVOUS_VERSION }}
version: v0.69.3
format: table
exit-code: "1"
ignore-unfixed: false
severity: HIGH,CRITICAL
- name: Generate container SPDX inventory
uses: aquasecurity/trivy-action@57a97c7e7821a5776cebc9bb87c984fa69cba8f1 # v0.35.0, post-incident safe SHA
with:
image-ref: git.finalfactory.de/heikyu/rendezvous:${{ env.RENDEZVOUS_VERSION }}
version: v0.69.3
format: spdx-json
output: ${{ env.RENDEZVOUS_RELEASE_DIR }}/FinalFactory.Rendezvous.Container.${{ env.RENDEZVOUS_VERSION }}.spdx.json
- name: Finalize checksums over the publish-ready candidate
shell: bash
run: |
set -euo pipefail
source_date_epoch="$(git show -s --format=%ct HEAD)"
docker run --rm \
--user "$(id -u):$(id -g)" \
--volume "$GITHUB_WORKSPACE:/source" \
--volume "${RUNNER_TEMP}:${RUNNER_TEMP}" \
--workdir /source \
"$RENDEZVOUS_RELEASE_BUILDER" \
bash -c 'python3 eng/release_artifacts.py normalize-container-sbom \
--file "$1/FinalFactory.Rendezvous.Container.$2.spdx.json" \
--version "$2" \
--commit "$3" \
--source-date-epoch "$4" \
&& ./scripts/finalize-release-candidate.sh "$2" "$1"' \
_ "$RENDEZVOUS_RELEASE_DIR" "$RENDEZVOUS_VERSION" "$GITHUB_SHA" "$source_date_epoch"
- name: Preserve verified candidate artifacts
uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
with:
name: rendezvous-${{ env.RENDEZVOUS_VERSION }}
path: ${{ env.RENDEZVOUS_RELEASE_DIR }}
if-no-files-found: error
retention-days: 30
- name: Install pinned signing client
uses: sigstore/cosign-installer@6f9f17788090df1f26f669e9d70d6ae9567deba6 # v4.1.2
with:
cosign-release: v3.0.6
- name: Publish once, sign, attest, and create release
shell: bash
env:
RENDEZVOUS_RELEASE_USERNAME: ${{ secrets.RELEASE_USERNAME }}
RENDEZVOUS_RELEASE_TOKEN: ${{ secrets.RELEASE_TOKEN }}
COSIGN_PRIVATE_KEY: ${{ secrets.COSIGN_PRIVATE_KEY }}
COSIGN_PASSWORD: ${{ secrets.COSIGN_PASSWORD }}
run: ./scripts/publish-release.sh "$RENDEZVOUS_VERSION" "$RENDEZVOUS_RELEASE_DIR"
+6
View File
@@ -6,3 +6,9 @@ TestResults/
*.suo *.suo
*.user *.user
*.userosscache *.userosscache
deploy/compose/.smoke.env
artifacts/
__pycache__/
*.pyc
deploy/compose/secrets/*
!deploy/compose/secrets/.gitignore
+25
View File
@@ -0,0 +1,25 @@
# Changelog
All notable Rendezvous release changes are recorded here. Versions follow
Semantic Versioning; HTTP, UDP, and connection-ticket format compatibility is
tracked separately and called out for every release.
## 1.0.0 - 2026-07-16
### Compatibility
- Initial Client and Contracts package major: 1.
- HTTP contract: v1; UDP mediation contract: v1; connection-ticket format: v1.
- Server accepts Client 1.0.0 through the latest compatible 1.x release.
- Client traversal is pinned to LiteNetLib 2.1.4; LiteNetLib 1.x is unsupported.
### Security and configuration
- Packages contain no reusable credentials or environment configuration.
- Production server startup requires provisioned signing keys and the hardened
single-active deployment configuration.
### Migration
- This is the first packaged release; no prior package or wire migration exists.
- Consumers must pin both Rendezvous packages to the same exact version.
+28
View File
@@ -1,4 +1,5 @@
<Project> <Project>
<Import Project="eng/Versions.props" />
<PropertyGroup> <PropertyGroup>
<AnalysisLevel>latest-recommended</AnalysisLevel> <AnalysisLevel>latest-recommended</AnalysisLevel>
<ContinuousIntegrationBuild Condition="'$(CI)' == 'true'">true</ContinuousIntegrationBuild> <ContinuousIntegrationBuild Condition="'$(CI)' == 'true'">true</ContinuousIntegrationBuild>
@@ -8,8 +9,35 @@
<ImplicitUsings>enable</ImplicitUsings> <ImplicitUsings>enable</ImplicitUsings>
<LangVersion>latest</LangVersion> <LangVersion>latest</LangVersion>
<Nullable>enable</Nullable> <Nullable>enable</Nullable>
<Version>$(RendezvousVersion)</Version>
<PackageVersion Condition="'$(PackageVersion)' == ''">$(RendezvousVersion)</PackageVersion>
<AssemblyVersion>$(RendezvousMajorVersion).0.0.0</AssemblyVersion>
<FileVersion>$(RendezvousMajorVersion).$(RendezvousMinorVersion).$(RendezvousPatchVersion).0</FileVersion>
<Authors>Final Factory</Authors>
<Company>Final Factory</Company>
<RepositoryUrl>https://git.finalfactory.de/HeiKyu/Rendezvous</RepositoryUrl>
<RepositoryType>git</RepositoryType>
<PackageProjectUrl>https://git.finalfactory.de/HeiKyu/Rendezvous</PackageProjectUrl>
<PublishRepositoryUrl>true</PublishRepositoryUrl>
<EmbedUntrackedSources>true</EmbedUntrackedSources>
<EnableSourceLink>true</EnableSourceLink>
<IncludeSymbols>true</IncludeSymbols>
<SymbolPackageFormat>snupkg</SymbolPackageFormat>
<PackageReleaseNotes>See CHANGELOG.md in the package and repository.</PackageReleaseNotes>
<RestorePackagesWithLockFile>true</RestorePackagesWithLockFile> <RestorePackagesWithLockFile>true</RestorePackagesWithLockFile>
<RestoreLockedMode Condition="'$(CI)' == 'true'">true</RestoreLockedMode> <RestoreLockedMode Condition="'$(CI)' == 'true'">true</RestoreLockedMode>
<NuGetAudit>true</NuGetAudit>
<NuGetAuditMode>all</NuGetAuditMode>
<NuGetAuditLevel>moderate</NuGetAuditLevel>
<TreatWarningsAsErrors>true</TreatWarningsAsErrors> <TreatWarningsAsErrors>true</TreatWarningsAsErrors>
</PropertyGroup> </PropertyGroup>
<Target Name="ConfigureGiteaSourceLink"
BeforeTargets="_GenerateSourceLinkFile"
DependsOnTargets="InitializeSourceControlInformation">
<ItemGroup>
<SourceRoot Update="@(SourceRoot)"
SourceLinkUrl="$(RepositoryUrl)/raw/commit/$(SourceRevisionId)/*" />
</ItemGroup>
</Target>
</Project> </Project>
+1 -1
View File
@@ -4,7 +4,7 @@
<CentralPackageTransitivePinningEnabled>true</CentralPackageTransitivePinningEnabled> <CentralPackageTransitivePinningEnabled>true</CentralPackageTransitivePinningEnabled>
</PropertyGroup> </PropertyGroup>
<ItemGroup> <ItemGroup>
<PackageVersion Include="LiteNetLib" Version="2.1.4" /> <PackageVersion Include="LiteNetLib" Version="[$(LiteNetLibVersion)]" />
<PackageVersion Include="Microsoft.AspNetCore.OpenApi" Version="10.0.9" /> <PackageVersion Include="Microsoft.AspNetCore.OpenApi" Version="10.0.9" />
<PackageVersion Include="Microsoft.Extensions.ApiDescription.Server" Version="10.0.9" /> <PackageVersion Include="Microsoft.Extensions.ApiDescription.Server" Version="10.0.9" />
<PackageVersion Include="Microsoft.NET.Test.Sdk" Version="18.4.0" /> <PackageVersion Include="Microsoft.NET.Test.Sdk" Version="18.4.0" />
+35
View File
@@ -0,0 +1,35 @@
# syntax=docker/dockerfile:1.7@sha256:a57df69d0ea827fb7266491f2813635de6f17269be881f696fbfdf2d83dda33e
FROM mcr.microsoft.com/dotnet/sdk:10.0.301-noble@sha256:ea8bde36c11b6e7eec2656d0e59101d4462f6bd630730f2c8201ed0572b295d5 AS build
ARG SOURCE_REVISION_ID
WORKDIR /source
COPY Directory.Build.props Directory.Packages.props NuGet.config global.json Rendezvous.slnx ./
COPY eng/Versions.props eng/Versions.props
COPY src/FinalFactory.Rendezvous.Contracts/FinalFactory.Rendezvous.Contracts.csproj src/FinalFactory.Rendezvous.Contracts/packages.lock.json src/FinalFactory.Rendezvous.Contracts/
COPY src/FinalFactory.Rendezvous.Server/FinalFactory.Rendezvous.Server.csproj src/FinalFactory.Rendezvous.Server/packages.lock.json src/FinalFactory.Rendezvous.Server/
RUN dotnet restore src/FinalFactory.Rendezvous.Server/FinalFactory.Rendezvous.Server.csproj --locked-mode
COPY src/FinalFactory.Rendezvous.Contracts/ src/FinalFactory.Rendezvous.Contracts/
COPY src/FinalFactory.Rendezvous.Server/ src/FinalFactory.Rendezvous.Server/
RUN dotnet publish src/FinalFactory.Rendezvous.Server/FinalFactory.Rendezvous.Server.csproj \
--configuration Release \
--no-restore \
--output /out \
/p:UseAppHost=false \
/p:OpenApiGenerateDocuments=false \
/p:ContinuousIntegrationBuild=true \
/p:RepositoryCommit="$SOURCE_REVISION_ID" \
/p:SourceRevisionId="$SOURCE_REVISION_ID"
FROM mcr.microsoft.com/dotnet/aspnet:10.0.9-noble-chiseled@sha256:f820c4fbfb8bb204c3bbe05c69d48cd039cd0e67aa8f13ac1cec168819b90643 AS runtime
ENV ASPNETCORE_HTTP_PORTS=8080 \
DOTNET_EnableDiagnostics=0 \
DOTNET_CLI_TELEMETRY_OPTOUT=1 \
TMPDIR=/tmp
WORKDIR /app
COPY --from=build --chown=1654:1654 /out/ ./
USER 1654:1654
EXPOSE 8080/tcp
EXPOSE 9050/udp
ENTRYPOINT ["dotnet", "FinalFactory.Rendezvous.Server.dll"]
+39 -6
View File
@@ -17,7 +17,7 @@ Rendezvous is intended to provide:
- Isolation by game, environment, protocol version, and region. - Isolation by game, environment, protocol version, and region.
- Operational health, metrics, logging, administration, and rate limiting. - Operational health, metrics, logging, administration, and rate limiting.
UDP hole punching cannot guarantee a direct connection through every network. Symmetric NAT, carrier-grade NAT, restrictive firewalls, and platform policies can prevent it. Consumers must therefore support a defined fallback, such as a dedicated server or a future relay service. UDP hole punching cannot guarantee a direct connection through every network. Symmetric NAT, carrier-grade NAT, restrictive firewalls, and platform policies can prevent it. Consumers must therefore support a defined fallback, such as a dedicated server. The v1 SDK returns an optional game-configured endpoint for an explicit caller decision; it never routes automatically, and v1 does not provide a relay.
## Connection flow ## Connection flow
@@ -38,7 +38,11 @@ UDP hole punching cannot guarantee a direct connection through every network. Sy
- `FinalFactory.Rendezvous.TestClient` — thin interactive and scriptable host/browser/join diagnostic built only on the public SDK. - `FinalFactory.Rendezvous.TestClient` — thin interactive and scriptable host/browser/join diagnostic built only on the public SDK.
- `FinalFactory.Rendezvous.Tests` — unit, integration, security, and connection-lifecycle tests. - `FinalFactory.Rendezvous.Tests` — unit, integration, security, and connection-lifecycle tests.
The server directory and NAT mediator begin as separate modules in one deployable service because they share session, lease, authorization, and endpoint state. Their internal boundary should allow independent deployment later if scale, availability, or security requirements diverge. The server directory and NAT mediator are separate modules in one single-active
deployable service because they share ephemeral session, lease, authorization,
replay, and endpoint state. Their internal boundary can support a future
explicitly designed shared-state architecture; operators must not create
multiple active v1 replicas.
## Service boundaries ## Service boundaries
@@ -75,7 +79,15 @@ The initial service does not provide:
## Project status ## Project status
Rendezvous is currently in its initial design and bootstrap stage. The first implementation should establish the contracts, directory leases, LiteNetLib mediator, client SDK, thin test client, and a three-party integration test before either game depends on it for production connectivity. Rendezvous is under active roadmap development. The versioned contracts,
directory leases, authenticated join attempts, LiteNetLib mediator, caller-owned
SDK coordination, typed connection outcomes, thin public-SDK diagnostic client,
deterministic NAT topology harness, hostile-input controls,
observability/operator surface, secure single-active Linux deployment, and
numeric capacity/resilience gates, and reproducible signed release pipeline are
implemented. Consumer pilots and final production-readiness gates remain in progress;
participating games must not treat the current repository as a finished production
service until those gates land.
The ratified v1 boundaries, trust decisions, privacy rules, safety budgets, and The ratified v1 boundaries, trust decisions, privacy rules, safety budgets, and
threat model are indexed in [the architecture documentation](docs/architecture/README.md). threat model are indexed in [the architecture documentation](docs/architecture/README.md).
@@ -83,6 +95,25 @@ The frozen v1 wire surface is documented in the
[HTTP, UDP, and generated OpenAPI contracts](docs/contracts/README.md). [HTTP, UDP, and generated OpenAPI contracts](docs/contracts/README.md).
Tenant policy, publisher/operator principals, and production key custody are Tenant policy, publisher/operator principals, and production key custody are
defined in [game provisioning and signing-key lifecycle](docs/security/provisioning.md). defined in [game provisioning and signing-key lifecycle](docs/security/provisioning.md).
Layered HTTP/UDP budgets, overload behavior, and safe operational tuning are
defined in [hostile-input and overload protection](docs/security/abuse-protection.md).
Health semantics, bounded telemetry, alerting, audit privacy, and the authenticated
operator controls are defined in the
[observability and operator runbook](docs/operations/observability-and-operator-runbook.md).
The pinned non-root container, production topology, graceful drain, Linux
hardening, smoke procedure, and recovery lifecycle are documented in
[secure single-active Linux deployment](docs/deployment/linux.md).
The numeric core-state candidate profile, public launch objectives, accelerated
soak, resilience matrix, and single-active scaling decision are recorded in
[capacity and resilience gates](docs/operations/capacity-and-resilience.md).
Release versions, compatibility windows, immutable artifact construction,
signing, staged promotion, rollback, and migration are defined in
[releases and compatibility](docs/releases/README.md).
The scriptable host/browser/join diagnostic and its stable automation contract are
documented in the [TestClient integration guide](docs/integration/test-client.md).
The always-on three-party scenarios, optional Linux namespace topology, and
simulation limits are documented in the
[deterministic topology harness](docs/integration/topology-harness.md).
## Development ## Development
@@ -99,9 +130,11 @@ dotnet test Rendezvous.slnx --configuration Release --no-build
Run the bootstrap server with Run the bootstrap server with
`dotnet run --project src/FinalFactory.Rendezvous.Server`. It serves HTTP health endpoints and binds `dotnet run --project src/FinalFactory.Rendezvous.Server`. It serves HTTP health endpoints and binds
the configured UDP mediator port; both stop through normal host cancellation. the configured UDP mediator port; both stop through normal host cancellation.
The launch profile uses an ephemeral development-only signing key. Production The launch profile uses separate ephemeral development-only publisher and operator
startup fails closed until externally supplied game policies and `env:` signing signing keys. Production
key references resolve to valid key material; no reusable game secret is stored startup fails closed until its advertised endpoints, proxy trust boundary,
externally supplied game policies, and `env:` (base64) or `file:` (raw,
absolute, non-symlink) signing-key references resolve safely; no reusable game secret is stored
in this repository or the public Client package. in this repository or the public Client package.
The project dependency rules and supported runtime choices are documented in The project dependency rules and supported runtime choices are documented in
[project and dependency boundaries](docs/architecture/project-boundaries.md). [project and dependency boundaries](docs/architecture/project-boundaries.md).
+1
View File
@@ -6,6 +6,7 @@
<Project Path="src/FinalFactory.Rendezvous.TestClient/FinalFactory.Rendezvous.TestClient.csproj" /> <Project Path="src/FinalFactory.Rendezvous.TestClient/FinalFactory.Rendezvous.TestClient.csproj" />
</Folder> </Folder>
<Folder Name="/tests/"> <Folder Name="/tests/">
<Project Path="tests/FinalFactory.Rendezvous.Capacity/FinalFactory.Rendezvous.Capacity.csproj" />
<Project Path="tests/FinalFactory.Rendezvous.Tests/FinalFactory.Rendezvous.Tests.csproj" /> <Project Path="tests/FinalFactory.Rendezvous.Tests/FinalFactory.Rendezvous.Tests.csproj" />
</Folder> </Folder>
</Solution> </Solution>
@@ -0,0 +1,58 @@
{
"AllowedHosts": "localhost;127.0.0.1",
"Rendezvous": {
"Deployment": {
"PublicHttpBaseUrl": "https://localhost/",
"PublicUdpHost": "127.0.0.1",
"PublicUdpPort": 9050,
"DrainDeadlineSeconds": 30,
"MinimumDrainSeconds": 1,
"SingleActiveInstance": true,
"AllowPrivatePublicEndpoints": true
},
"Udp": {
"ListenAddress": "0.0.0.0",
"Port": 9050
},
"AbuseProtection": {
"TrustedProxyAddresses": ["127.0.0.1"],
"OperatorAllowedAddresses": ["127.0.0.1"]
},
"Provisioning": {
"Issuer": "final-factory-rendezvous-smoke",
"Audience": "rendezvous-service",
"ClockSkewSeconds": 30,
"SigningKeys": [
{
"KeyId": "local-smoke-1",
"SecretReference": "file:/run/secrets/rendezvous-signing-key",
"CredentialKinds": ["DedicatedPublisher"],
"GameId": "space-game",
"EnvironmentId": "smoke",
"NotBefore": "2026-01-01T00:00:00Z",
"SignUntil": "2100-01-01T00:00:00Z",
"VerifyUntil": "2100-01-02T00:00:00Z"
}
],
"Games": [
{
"GameId": "space-game",
"EnvironmentId": "smoke",
"Enabled": true,
"ProtocolVersions": [1],
"Regions": ["local"],
"VisibilityModes": ["Public"],
"PublisherTrustModes": ["ManagedDedicated"],
"MetadataValueMaxBytes": {},
"RequiredMetadataKeys": [],
"MetadataMaxBytes": 512,
"MetadataMaxKeys": 0,
"MaxListingsPerPrincipal": 10,
"MaxAnonymousListingsPerAddress": 0,
"MaxActiveJoinAttempts": 100,
"FallbackPolicy": "Disabled"
}
]
}
}
}
+35
View File
@@ -0,0 +1,35 @@
name: rendezvous-local
services:
rendezvous:
image: "${RENDEZVOUS_IMAGE:-finalfactory/rendezvous:local}"
build:
context: ../..
dockerfile: Dockerfile
init: true
user: "${RENDEZVOUS_UID:?set RENDEZVOUS_UID to a non-root host UID}:${RENDEZVOUS_GID:?set RENDEZVOUS_GID to its GID}"
read_only: true
tmpfs:
- /tmp:rw,noexec,nosuid,nodev,size=16m,uid=${RENDEZVOUS_UID},gid=${RENDEZVOUS_GID},mode=0700
cap_drop:
- ALL
security_opt:
- no-new-privileges:true
pids_limit: 128
mem_limit: 512m
cpus: 1.0
ulimits:
nofile:
soft: 4096
hard: 4096
stop_grace_period: 40s
restart: unless-stopped
environment:
ASPNETCORE_ENVIRONMENT: Production
ASPNETCORE_HTTP_PORTS: "8080"
volumes:
- ./appsettings.Production.json:/app/appsettings.Production.json:ro
- ./secrets/signing-key:/run/secrets/rendezvous-signing-key:ro
ports:
- "127.0.0.1:8080:8080/tcp"
- "9050:9050/udp"
+2
View File
@@ -0,0 +1,2 @@
*
!.gitignore
+48
View File
@@ -0,0 +1,48 @@
[Unit]
Description=Final Factory Rendezvous service
Documentation=https://git.finalfactory.de/HeiKyu/Rendezvous
After=network-online.target time-sync.target
Wants=network-online.target time-sync.target
[Service]
Type=simple
User=rendezvous
Group=rendezvous
WorkingDirectory=/opt/rendezvous
ExecStart=/usr/bin/dotnet /opt/rendezvous/FinalFactory.Rendezvous.Server.dll
Environment=ASPNETCORE_ENVIRONMENT=Production
Environment=ASPNETCORE_HTTP_PORTS=8080
Environment=DOTNET_EnableDiagnostics=0
EnvironmentFile=-/etc/rendezvous/rendezvous.env
Restart=on-failure
RestartSec=5s
KillSignal=SIGTERM
KillMode=mixed
TimeoutStopSec=40s
NoNewPrivileges=true
PrivateDevices=true
PrivateTmp=true
ProtectClock=true
ProtectControlGroups=true
ProtectHome=true
ProtectHostname=true
ProtectKernelLogs=true
ProtectKernelModules=true
ProtectKernelTunables=true
ProtectSystem=strict
RestrictAddressFamilies=AF_UNIX AF_INET AF_INET6
RestrictNamespaces=true
RestrictRealtime=true
RestrictSUIDSGID=true
CapabilityBoundingSet=
AmbientCapabilities=
LockPersonality=true
SystemCallArchitectures=native
UMask=0077
LimitNOFILE=4096
CPUQuota=200%
MemoryMax=2G
TasksMax=128
[Install]
WantedBy=multi-user.target
File diff suppressed because it is too large Load Diff
@@ -64,7 +64,7 @@ them but must not raise them without security review.
| Browser page | 100 listings and 256 KiB encoded response; opaque cursor; stable bounded sort | | Browser page | 100 listings and 256 KiB encoded response; opaque cursor; stable bounded sort |
| UDP datagram accepted | 1,200 bytes; oversized or fragmented application payloads are dropped without response | | UDP datagram accepted | 1,200 bytes; oversized or fragmented application payloads are dropped without response |
| Opaque HTTP credential | 1,024 bytes encoded | | Opaque HTTP credential | 1,024 bytes encoded |
| UDP capability or ticket | 768 bytes encoded, with the complete datagram still at most 1,200 bytes | | UDP capability or connection ticket | 192 base64url characters; NAT punch capabilities also remain below LiteNetLib's 256-character token ceiling; complete datagram at most 1,200 bytes |
| Clock skew | 30 seconds maximum when validating issued/not-before/expiry times | | Clock skew | 30 seconds maximum when validating issued/not-before/expiry times |
| Lease lifetime | 60 seconds; renewal accepted from 30 seconds; no client-selected extension | | Lease lifetime | 60 seconds; renewal accepted from 30 seconds; no client-selected extension |
| Host presence freshness | 20 seconds | | Host presence freshness | 20 seconds |
@@ -129,9 +129,18 @@ until the owner records:
- which games may enable anonymous unlisted player hosting; - which games may enable anonymous unlisted player hosting;
- deployment regions, data-processing jurisdiction, and approval of the stated - deployment regions, data-processing jurisdiction, and approval of the stated
30-day audit/13-month aggregate retention periods; 30-day audit/13-month aggregate retention periods;
- the per-game dedicated fallback endpoint policy; - the per-game dedicated fallback endpoint policy.
- the measured supported profile and whether the 99.5% single-active objective
is sufficient or shared-state/high-availability work must be brought forward. Issue #18 measured and ratified the original 2-vCPU/2-GiB, 25,000-listing,
10,000-attempt core-state candidate profile and retained the 99.5% single-active
topology. It does not claim that core measurements prove public HTTP/UDP SLOs.
The versioned evidence, RTO, failure domains, and explicit signals that trigger
shared-state/high-availability work are recorded in the
[capacity and resilience gate](../operations/capacity-and-resilience.md). The
real-network canary in #23 must confirm that the proposed regional launch load
fits this profile and validate the public SLOs; it may lower the launch cap but
may not silently enable a
second active instance.
These are configuration and launch decisions, not permission to weaken the These are configuration and launch decisions, not permission to weaken the
tenant, replay, endpoint-verification, or secret-handling controls. tenant, replay, endpoint-verification, or secret-handling controls.
@@ -60,12 +60,14 @@ the supplied ID.
### UDP presence ### UDP presence
Only a structurally valid `HostPresence` datagram with the issued capability can Only a structurally valid frozen `HostPresence` envelope or native LiteNetLib
refresh presence. The public endpoint is the UDP packet's observed source on the host-presence request with the issued capability can refresh presence. The public
host's gameplay socket; the HTTP API never accepts one. The bounded local candidate endpoint is the UDP packet's observed source on the host's gameplay socket; the
comes from the authenticated datagram. Invalid, unknown, or client-presence packets HTTP API never accepts one. The bounded local candidate comes from the authenticated
receive no response. Presence expiry demotes public visibility but keeps the lease, packet. Invalid or unknown inputs receive no response. ADR 0009 defines the later
so the same handle can restore visibility without changing session identity. attempt-role use of frozen `ClientPresence` and native host/client requests.
Presence expiry demotes public visibility but keeps the lease, so the same handle
can restore visibility without changing session identity.
Public listing responses contain bounded listing data only. They never contain Public listing responses contain bounded listing data only. They never contain
public/local endpoints, lease tokens, presence capabilities, fingerprints, store public/local endpoints, lease tokens, presence capabilities, fingerprints, store
@@ -0,0 +1,78 @@
# ADR 0008: scoped join attempts and one-time connection tickets
- Status: Accepted
- Date: 2026-07-16
- Tracking: #10
## Decision
Join creation is an unauthenticated public operation because v1 does not treat a
Rendezvous caller as game identity. The HTTP source address is normalized and
converted to a process-keyed opaque subject for idempotency and bounded policy
accounting; raw addresses and the derived subject are never returned or logged.
A successful request means only that this network client may try to connect to
this active session. It does not reserve capacity or grant gameplay admission.
Creation validates the v1 contract, caller idempotency key, enabled tenant policy,
exact gameplay protocol, listing scope, live lease, and fresh authenticated host
presence in one atomic store operation. A listing advertised as full remains
joinable because its player count is advisory and the game host owns the final
capacity, identity, ban, and admission decision.
Each attempt derives independent host-punch, client-punch, and connection-ticket
credentials plus opaque attempt and mediation IDs from a process-ephemeral HMAC
key, the client subject, the complete canonical request fingerprint, a fresh salt,
and a purpose/role label. Credentials are 32-byte base64url values (43 characters),
below both the 192-character Rendezvous capability ceiling and LiteNetLib's
256-character NAT token ceiling. The connection ticket uses half of that payload
for its attempt ID and half for an independently derived 128-bit authenticator, so
the SDK can correlate concurrent introductions without increasing UDP response
size. State retains keyed credential fingerprints, derivation inputs, and salt—not
issued plaintext. All diagnostic string representations redact credentials and
derivation material.
The client receives only its punch capability. A host polls its own listing with
the lease token in `X-Rendezvous-Lease-Token` and receives only host-role
capabilities through a signed, listing-bound, five-minute cursor. Replaying an
identical join request returns the same live attempt; changing the request under
the same owner/key conflicts. A client may cancel with its punch capability in
`X-Rendezvous-Client-Punch-Capability`; cancellation atomically marks the attempt
and retains a bounded tombstone until its original expiry. Host polling returns
that tombstone so a coordinator can revoke any local ticket authorization, while
endpoint binding, introduction, ticket issuance, and ticket consumption all
reject the cancelled attempt. Listing deletion, expiry, revocation, or process
restart removes every associated attempt and credential fingerprint.
Endpoint binding remains role- and capability-specific. The first endpoint
observed for a role wins atomically; an exact UDP duplicate is idempotent, while
endpoint or role substitution is rejected. An introduction is consumable once
only after both roles bind, so concurrent attempts for the same listing cannot
cross-wire.
The connection ticket is distinct from both punch capabilities and is reproduced
only after introduction succeeds. Its window begins at that moment and lasts at
most 20 seconds without outliving the 30-second attempt. The server has an atomic
fingerprint-consumption seam for mediator tests and revocation. On the game host,
the SDK's bounded `ConnectionTicketValidator` stores a process-keyed digest,
accepts an exact ticket once under a lock, rejects altered/cross-attempt/expired/
revoked/replayed tickets, and zeroes retained digests and key material on disposal.
Issue #11 carries the fixed-size ticket in the authenticated introduction. Issue
#12 extracts its embedded attempt ID, bounds the host's local authorization window
by both the host-polled attempt expiry and the configured ticket lifetime, then
wires one-time consumption into the caller-owned coordinator. Both peers receive
a digest of the exact expected ticket over HTTP and reject any syntactically valid
but unauthenticated introduction token. Embedding the ID prevents concurrent or
late introductions from cross-binding a valid ticket while preserving the
mediator's 2.0 response-byte amplification ceiling.
## Consequences
- A join attempt is transport authorization, never proof of player identity or a
game slot.
- Network-address-derived subjects are process-local abuse/idempotency scopes,
not stable user identifiers; stronger authenticated player scopes require a
future game-owned identity contract.
- Cancellation after a ticket has reached a host must also revoke that host's
local validator entry; coordinator wiring owns that race in issue #12.
- Capability and ticket plaintext never enter browser results, state snapshots,
logs, metrics, or generated string representations.
@@ -0,0 +1,68 @@
# ADR 0009: authenticated bounded LiteNetLib NAT mediator
- Status: Accepted
- Date: 2026-07-16
- Tracking: #11
## Decision
The server owns one LiteNetLib `NetManager` and its `NatPunchModule` on the
configured UDP endpoint. It runs in manual mode with a configured maximum number
of datagrams per poll and a short caller-owned poll interval. LiteNetLib events
are unsynchronized so authenticated requests are processed immediately on that
single polling path rather than accumulated in an unbounded event queue. The
mediator never accepts a LiteNetLib gameplay connection or handles application
payloads.
The packet layer also consumes the frozen v1 presence envelope on the same
socket. Native NAT requests use a canonical fixed-size 192-character token that
binds a role (`HostPresence`, attempt `Host`, or attempt `Client`), mediation
handle, and the already-issued capability. Both transports enter one processor
and the same atomic store operations. No transport-supplied public address is
trusted; the socket source is authoritative.
LiteNetLib's native NAT packet family also contains introduction-response and
punch frames that are appropriate for peers but unsafe on a public mediator: a
forged response can name arbitrary destinations. The packet layer therefore
decodes only the pinned `NatIntroduceRequest` wire shape and consumes every
inbound packet before `NatPunchModule` sees it. The module is outbound-only and
may send introductions solely from a completed authorized plan.
Listing presence refreshes authorize no response. Attempt contributions bind the
first observed endpoint for exactly one capability role. Exact duplicates are
idempotent; a different endpoint, the opposite role, an expired/cancelled
attempt, or a stale listing presence cannot replace it. The introduction is
consumed atomically only after both roles bind and their observed address
families match, preventing concurrent attempts for one listing from cross-wiring.
A reported local candidate is eligible only when it is RFC 1918 IPv4 or IPv6
unique-local unicast, matches the observed family, and both peers have the same
observed public address. Otherwise `NatIntroduce` receives the observed public
endpoint in the local slot. Loopback, link-local, multicast, unspecified,
documentation IPv6, global-address claims, and cross-family claims are never
disclosed as local targets. IPv4 is required; observed global IPv6 can be used
when both peers contribute IPv6, without claiming guaranteed IPv6 NAT traversal.
The introduction carries only the distinct connection ticket and is emitted at
most once to each verified observed endpoint. The fixed authenticated native
request and bounded frozen envelope keep the combined response bytes within the
2.0 verified amplification budget; unauthenticated inputs receive zero bytes.
Malformed, truncated, oversized, spoofed, or unrelated LiteNetLib packets do not
grow Rendezvous state. Raw endpoints and credentials are never logged or exposed
through diagnostic string representations.
Frozen IPv6 listing-presence refresh remains valid because it emits no response.
IPv6 attempt roles require the fixed-size native LiteNetLib request; accepting the
short frozen envelope would exceed the 2.0 byte budget for two IPv6 introduction
frames. The required IPv4 listen address and optional IPv6 listen address are
configured separately so enabling one family never widens the other family to a
wildcard bind.
## Consequences
- Hosts refresh listing presence and answer invitations from their actual
gameplay socket; a separate mediator socket would observe the wrong mapping.
- Caller-owned SDK coordination in #12 must poll the host invitation endpoint,
send the corresponding native role token, and consume the returned ticket.
- UDP loss can prevent traversal, but it cannot cause an arbitrary destination,
replay, role substitution, or cross-attempt introduction.
@@ -0,0 +1,117 @@
# ADR 0010: typed connection outcomes, deadlines, and caller-owned fallback
- Status: Accepted
- Date: 2026-07-16
- Tracking: #13
## Context
A connection can stop in the directory, authorization, mediation, NAT traversal,
or direct-connection phase. Those failures have different authorities: an HTTP
response can authoritatively reject a join, the SDK can observe a local timeout,
and only the remote host can reject a direct connection. Treating all of them as
one message or generic timeout would make player guidance, retry policy, tests,
and operational measurements unreliable.
UDP loss, service silence, cancellation, and late LiteNetLib callbacks also make
completion races unavoidable. Games need one terminal result and bounded work,
not a sequence of contradictory callbacks. Direct traversal cannot be guaranteed,
but v1 has no gameplay relay and must not imply otherwise.
## Decision
### Closed typed outcome model
`ConnectionOutcomeKind` is the stable wire-level terminal set: connected,
cancelled, directory not found, attempt expired, incompatible protocol,
unauthorized, rate limited, no host presence, service unavailable or rejected,
mediator unavailable, punch timeout, direct-connect timeout, host rejection,
transport error, manager stopped, and disposed.
The already-frozen v1 members `TimedOut`, `StaleHost`, `TransportFailed`, and
`FallbackOffered` retain their original numeric values for source and wire
compatibility. New SDK code never emits them. The report service accepts them,
normalizes the first three to their precise modern equivalents, and does not let
legacy compatibility weaken the typed coordinator result.
The client adds `RendezvousConnectionOutcomeSource`, failure category, and phase.
These fields preserve authority instead of guessing from text:
- `RendezvousService` is used only for an HTTP decision or bounded service
silence. Its optional `ServiceError` retains the stable service error code.
- `LocalTraversal` reports local punch, direct-connect, and transport
observations.
- `RemoteHost` reports an explicit direct-connection rejection.
- `Caller` and `Lifecycle` distinguish cancellation from manager shutdown or
disposal.
Messages remain diagnostic and are never parsed into outcomes. A successful NAT
introduction is only a transition to direct connection; `Connected` is emitted
only after LiteNetLib reports the authenticated peer connected.
Join issuance is exposed as `RendezvousConnectionStartResult`, containing exactly
one issued attempt or one terminal service outcome. Once an attempt is issued,
the coordinator owns its local terminal outcome. Completion is exactly once;
terminal paths release SDK subscriptions so late introductions, peer callbacks,
network errors, cancellation, and polling are inert.
### Bounded phases and retries
Each HTTP try has a five-second default silence budget, configurable from above
zero through thirty seconds. Only safe operations use the existing bounded retry
policy, honoring caller cancellation and server retry guidance. Exhausting that
budget returns `ServiceUnavailable`; it never waits indefinitely.
Traversal has independent defaults: ten seconds for punch/mediation and five
seconds for the direct connection. Both are configurable up to thirty seconds.
Local budgets, retry schedules, and elapsed duration use monotonic time, so a
wall-clock correction cannot extend them or produce a negative duration. The
signed attempt expiry is converted to an additional monotonic upper bound when
the attempt is received. Punch retries retain
their bounded request count and exponential backoff; crossing a phase deadline
completes exactly once even if a delayed packet later arrives. Tests use an
injected clock and do not depend on wall-clock sleeps.
### Explicit dedicated fallback handoff
A publisher may attach one validated dedicated endpoint to registration or
update only when the tenant's provisioned fallback policy allows it. The server
copies that endpoint into browser and issued-attempt contracts.
The client coordinator defensively copies it into every terminal outcome; a game
may override it locally through `DedicatedFallbackOverride`.
The SDK never opens, dials, reserves, probes, or authenticates the fallback. The
game decides whether the outcome permits fallback, presents any player choice,
and connects through its own gameplay transport and admission rules. Absence of
an endpoint is an honest no-fallback result. Gameplay relay is absent from v1.
### Privacy-safe optional reporting
After an issued attempt completes, the game may explicitly report its outcome
with the short-lived client punch capability. Reporting is authenticated and
idempotent: an exact repeat succeeds as a duplicate, while a conflicting repeat
is rejected. Reports contain only an allowlisted outcome enum and one coarse
elapsed bucket (`<1s`, `15s`, `515s`, `1530s`, or `30s+`). They contain no
diagnostic message, exact duration, endpoint, metadata, player identifier, or
credential.
Frozen v1 DTOs still expose `elapsedMilliseconds` and `diagnosticCode`. They are
deprecated compatibility inputs: the current SDK omits them, the service
immediately buckets legacy elapsed time, and neither exact timing nor diagnostic
text is retained, logged, or used as a metric dimension.
The store retains a bounded capability-fingerprint tombstone long enough to
accept a report after the live attempt expires. Metrics count the first accepted
outcome only and use only outcome plus elapsed bucket as dimensions. Service
issuance failures cannot be reported because no attempt capability was issued.
## Consequences
- Player-facing UI can map stable outcome/category pairs to localized guidance
without exposing diagnostic strings.
- Service rejection, remote-host rejection, and local observation remain
distinguishable for retry and support decisions.
- Games own fallback policy and gameplay admission; Rendezvous does not claim a
guaranteed connection path.
- Outcome additions are contract changes and require OpenAPI, serialization,
public API, fake-clock, late-event, and idempotency coverage.
+3
View File
@@ -10,6 +10,9 @@ decision requires a superseding ADR and corresponding contract/test updates.
- [ADR 0005: authenticated session lease and presence lifecycle](0005-session-lease-lifecycle.md) - [ADR 0005: authenticated session lease and presence lifecycle](0005-session-lease-lifecycle.md)
- [ADR 0006: bounded compatible session browser](0006-compatible-session-browser.md) - [ADR 0006: bounded compatible session browser](0006-compatible-session-browser.md)
- [ADR 0007: caller-owned .NET publisher and browser SDK](0007-caller-owned-dotnet-client-sdk.md) - [ADR 0007: caller-owned .NET publisher and browser SDK](0007-caller-owned-dotnet-client-sdk.md)
- [ADR 0008: scoped join attempts and one-time connection tickets](0008-scoped-join-attempts-and-tickets.md)
- [ADR 0009: authenticated bounded LiteNetLib NAT mediator](0009-authenticated-litenet-nat-mediator.md)
- [ADR 0010: typed connection outcomes, deadlines, and caller-owned fallback](0010-typed-connection-outcomes-and-fallback.md)
- [Threat model](../security/threat-model.md) - [Threat model](../security/threat-model.md)
- [Security promise and test matrix](../security/control-matrix.md) - [Security promise and test matrix](../security/control-matrix.md)
- [Versioned HTTP and UDP contracts](../contracts/README.md) - [Versioned HTTP and UDP contracts](../contracts/README.md)
+4 -2
View File
@@ -14,5 +14,7 @@ The public .NET types live in `FinalFactory.Rendezvous.Contracts`, target
vectors and a public-API snapshot make accidental wire or source compatibility vectors and a public-API snapshot make accidental wire or source compatibility
changes fail the normal test gate. changes fail the normal test gate.
Any incompatible change requires a new contract version. Additive JSON fields Readers ignore unknown JSON members, but the release gate deliberately treats
may be introduced within v1 because v1 readers ignore unknown object members. any accepted OpenAPI or golden JSON surface drift as a contract-version change.
That conservative policy makes additive and incompatible published changes
equally visible to consumers instead of relying on an undocumented minor shape.
+27 -5
View File
@@ -33,15 +33,14 @@ the same value as a required query parameter.
| `GET` | `/v1/sessions` | Browse compatible public sessions. | | `GET` | `/v1/sessions` | Browse compatible public sessions. |
| `GET` | `/v1/sessions/{listingId}` | Resolve a public or explicitly shared unlisted listing. | | `GET` | `/v1/sessions/{listingId}` | Resolve a public or explicitly shared unlisted listing. |
| `POST` | `/v1/join-attempts` | Authorize and create a short-lived join attempt. | | `POST` | `/v1/join-attempts` | Authorize and create a short-lived join attempt. |
| `DELETE` | `/v1/join-attempts/{attemptId}` | Cancel an attempt using its client punch capability. |
| `GET` | `/v1/sessions/{listingId}/join-attempts` | Let an authenticated host poll pending attempts. | | `GET` | `/v1/sessions/{listingId}/join-attempts` | Let an authenticated host poll pending attempts. |
| `POST` | `/v1/join-attempts/{attemptId}/outcome` | Report a bounded connection outcome. | | `POST` | `/v1/join-attempts/{attemptId}/outcome` | Report a bounded connection outcome. |
| `GET` | `/health/live` | Report that the HTTP process is alive. | | `GET` | `/health/live` | Report that the HTTP process is alive. |
| `GET` | `/health/ready` | Report whether the UDP mediator is bound and ready. | | `GET` | `/health/ready` | Report whether the UDP mediator is bound and ready. |
The generated [OpenAPI document](../api/rendezvous-v1.json) is the normative The generated [OpenAPI document](../api/rendezvous-v1.json) is the normative
shape reference for parameters, bodies, and responses. Contract-only endpoints shape reference for parameters, bodies, and responses.
return `501` until their behavior is implemented by the subsequent directory,
lease, and join-orchestration issues.
Host polling sends its reusable lease credential in Host polling sends its reusable lease credential in
`X-Rendezvous-Lease-Token`; it must never be placed in a URL. Lease credentials `X-Rendezvous-Lease-Token`; it must never be placed in a URL. Lease credentials
@@ -49,6 +48,29 @@ for mutation operations are carried in their request bodies. Public browser
responses contain no IP endpoints, lease tokens, punch capabilities, connection responses contain no IP endpoints, lease tokens, punch capabilities, connection
tickets, player identifiers, or gameplay state. tickets, player identifiers, or gameplay state.
Attempt cancellation sends the short-lived client punch capability in
`X-Rendezvous-Client-Punch-Capability`. Join creation uses the observed HTTP
source only for a process-keyed, short-lived idempotency/abuse scope; this is not
player authentication and is never returned to callers.
Outcome reporting uses that same short-lived capability. It accepts only outcomes
for an issued attempt and carries one stable outcome enum plus one coarse elapsed
bucket. Exact duplicate reports are idempotent; conflicting repeats fail. Reports
never carry exact timing, diagnostics, endpoints, metadata, player identifiers,
or credentials.
The frozen v1 .NET request also retains deprecated `elapsedMilliseconds` and
`diagnosticCode` properties for source/wire compatibility. Current clients omit
them. If a legacy client supplies them, the server immediately converts elapsed
milliseconds to the coarse bucket and discards diagnostic text; neither value is
retained or used as a metric dimension.
Registration and update may include one validated `dedicatedFallback`. The
endpoint must be enabled by the tenant's provisioned fallback policy, is visible
browser data, and is copied into subsequently issued attempts.
It is a handoff for caller-owned policy: neither the HTTP service nor the SDK
automatically connects to it. V1 provides no gameplay relay.
## Idempotency, cursors, and retries ## Idempotency, cursors, and retries
Registration and join creation require a caller-generated visible-ASCII Registration and join creation require a caller-generated visible-ASCII
@@ -95,9 +117,9 @@ must not be parsed. Secrets and raw credentials are never echoed.
| 401 | `authenticationRequired` | | 401 | `authenticationRequired` |
| 403 | `forbidden` | | 403 | `forbidden` |
| 404 | `notFound` | | 404 | `notFound` |
| 409 | `conflict`, `incompatibleProtocol`, `replayRejected`, `capacityExceeded` | | 409 | `conflict`, `incompatibleProtocol`, `replayRejected` |
| 410 | `expired`, `staleHost` | | 410 | `expired`, `staleHost` |
| 429 | `rateLimited` (with retry guidance when known) | | 429 | `rateLimited`, `capacityExceeded` (with retry guidance when known) |
| 503 | `serviceUnavailable` (with retry guidance when known) | | 503 | `serviceUnavailable` (with retry guidance when known) |
| 500 | `internalError` | | 500 | `internalError` |
+51 -8
View File
@@ -1,11 +1,11 @@
# UDP presence contract v1 # UDP presence and NAT-punch contract v1
Tracking: #4 Tracking: #4, #11
The UDP mediator accepts a single bounded presence envelope from a host or The UDP mediator accepts the frozen bounded presence envelope below and native
client. It associates the authenticated mediation handle with the packet's LiteNetLib NAT-introduction requests. Both forms associate an authenticated
observed public source endpoint and the sender's reported local endpoint. It mediation handle with the packet's observed public source endpoint and the
does not carry gameplay packets. sender's reported local endpoint. Neither form carries gameplay packets.
All multi-byte integers use network byte order. UUID bytes use the canonical All multi-byte integers use network byte order. UUID bytes use the canonical
RFC 4122 textual order (the byte pairs from the 32 hexadecimal digits), not the RFC 4122 textual order (the byte pairs from the 32 hexadecimal digits), not the
@@ -49,5 +49,48 @@ Capabilities are short-lived, single-purpose, scoped to one mediation handle,
and compared without exposing them in logs. A valid-looking packet does not and compared without exposing them in logs. A valid-looking packet does not
prove authorization until the capability is checked. Invalid packets receive prove authorization until the capability is checked. Invalid packets receive
no UDP response, preventing the mediator from becoming an amplification oracle. no UDP response, preventing the mediator from becoming an amplification oracle.
Replay, expiry, pairing, and rate-limit policy are defined by later mediator For the frozen envelope, `HostPresence` is resolved against either the listing's
issues; the v1 envelope deliberately leaves no unbounded or reflected payload. host-presence capability or an attempt's host-role capability. `ClientPresence`
is resolved only against the attempt's client-role capability. Handles are
globally distinct in the active store, so this does not permit role confusion.
## Native LiteNetLib request token
A game using LiteNetLib sends `NatPunchModule.SendNatIntroduceRequest` from its
gameplay `NetManager`. The `additionalInfo` value is produced by
`NatPunchRequestTokenCodec` and is exactly 192 ASCII characters:
```text
rv1:<role>:<32 lowercase handle hex>:<43-character capability><dot padding>
```
`role` is `p` for listing host-presence refresh, `h` for the host side of a join
attempt, or `c` for its client side. Padding is canonical and leaves the token
below LiteNetLib's 256-character ceiling. Its fixed size also ensures that the
two authenticated introduction responses remain within the 2.0 response-byte
budget. Tokens with a wrong length, role, handle, capability, or padding receive
no response.
The mediator runs LiteNetLib in bounded manual-poll mode. Its packet layer admits
only the pinned native `NatIntroduceRequest` frame, consumes every inbound frame
before LiteNetLib can act on it, and uses `NatPunchModule` only to emit authorized
introductions. Native and frozen v1 inputs reach the same atomic role/capability
checks. Only the packet source is
used as the public endpoint. A claimed private candidate is retained only when
it is private unicast, matches the observed address family, and both authorized
peers were observed behind the same public address; otherwise the observed
public endpoint is substituted. IPv4 punching is required. IPv6 sources must be
observed global unicast and both roles must use IPv6; IPv6 NAT traversal remains
best-effort rather than a v1 release requirement.
The second valid contribution atomically consumes the introduction and starts
the connection-ticket lifetime. `NatIntroduce` is called once with the distinct
43-character connection ticket. Reordered and exact duplicate requests are
idempotent. Endpoint substitution, cross-role use, stale host presence, expired
or cancelled attempts, malformed packets, and gameplay payloads produce no
introduction and create no mediator queue or endpoint state.
Frozen envelopes may refresh listing presence over IPv6 because that operation
has no response. IPv6 attempt contributions must use the fixed-size native token;
the shorter frozen IPv6 envelope cannot fund two IPv6 introduction frames within
the 2.0 response-byte ceiling and is therefore dropped without response.
+228
View File
@@ -0,0 +1,228 @@
# Secure single-active Linux deployment
Tracking: #17
Rendezvous v1 stores listings, observed endpoints, join attempts, replay markers,
and runtime revocations only in the process that accepted them. Deploy exactly
one active instance. A second live replica would have a different directory and
replay boundary; `SingleActiveInstance=false` is therefore rejected rather than
presented as high availability.
## Pinned container
The root `Dockerfile` uses a multi-stage .NET 10 build and pins both Microsoft
base images by multi-architecture manifest digest. The runtime is the chiseled
ASP.NET image, contains only the published server, runs as UID/GID 1654, exposes
TCP 8080 and UDP 9050 explicitly, and does not require a writable application
directory. Supply a small writable `/tmp` tmpfs because runtime libraries can
legitimately need temporary space; keep the root filesystem read-only.
From a clean checkout:
```bash
docker build --pull=false --tag finalfactory/rendezvous:local .
docker inspect --format '{{.Config.User}}' finalfactory/rendezvous:local
```
The reported user must be `1654:1654`. Digest pins make a rebuild reproducible;
updating .NET is an explicit reviewed change to the tag, digest, SDK pin, and
lock files together. Do not replace the digest with `latest` in production.
The local Compose example applies a read-only root, non-root user, no Linux
capabilities, `no-new-privileges`, bounded PIDs/files/memory/CPU, and a shutdown
grace period longer than the service drain deadline:
```bash
install -d -m 0700 deploy/compose/secrets
umask 077
openssl rand -out deploy/compose/secrets/signing-key 32
export RENDEZVOUS_UID="$(id -u)"
export RENDEZVOUS_GID="$(id -g)"
test "$RENDEZVOUS_UID" -ne 0
docker compose -f deploy/compose/compose.yaml up --build --detach
```
`deploy/compose/appsettings.Production.json` is an isolated loopback smoke
profile, not an Internet template: it deliberately opts into private advertised
endpoints and has no TLS proxy. Its random key is ignored by Git and must be
deleted after use. Its deliberately long key window only keeps this disposable
local fixture usable; production keys require short, reviewed rotation windows.
Production configuration must use its real public names and must leave
`AllowPrivatePublicEndpoints` false.
## Production topology
Use one active service behind a source-preserving edge:
```text
clients -- HTTPS/443 --> TLS reverse proxy -- HTTP/8080 --> Rendezvous
clients -- UDP/9050 -------------------------------------> Rendezvous
```
- Give the HTTPS origin and UDP endpoint stable DNS names. Set
`PublicHttpBaseUrl` to the exact external HTTPS origin and `PublicUdpHost` /
`PublicUdpPort` to the endpoint given to game clients.
- Terminate TLS 1.2 or newer at a maintained reverse proxy. Bind internal HTTP
only to the private proxy network. Restrict `AllowedHosts` to the public HTTP
host; wildcard host filtering is rejected.
- Put only the proxy's exact literal addresses in
`Rendezvous:AbuseProtection:TrustedProxyAddresses`. Rendezvous ignores
forwarded headers from every other source. Keep the last proxy from replacing
the original client address and prevent direct access to TCP 8080.
- Forward UDP as UDP, without an HTTP proxy. NAT, load balancer, firewall, and
return routing must preserve the client's source IP/port and must send replies
from the same advertised IP/port. Many HTTP load balancers, Kubernetes ingress
controllers, rootless container port proxies, anycast products, and generic
L7 services cannot guarantee this. Do not deploy through one unless an actual
host/join smoke proves both observed source and reply path. A load balancer
must have exactly one healthy Rendezvous target.
- Permit inbound TCP 443 to the TLS proxy and UDP 9050 to Rendezvous. Permit the
proxy to reach TCP 8080. Permit DNS, time synchronization, image/telemetry
destinations as required by local policy, and UDP replies to client endpoints.
Deny public TCP 8080 and every unused inbound port.
Readiness is the load-balancer gate; liveness is only a process-health signal.
Remove a draining instance from new traffic when `/health/ready` becomes 503.
Do not use liveness failure to start a second active process while the old one
still owns the public UDP address.
## Required production configuration
Production startup validates all of these before binding listeners:
- an absolute path-free HTTPS `PublicHttpBaseUrl`;
- an unambiguous public `PublicUdpHost` and port;
- `SingleActiveInstance=true`, an explicit non-wildcard `AllowedHosts`, and at
least one exact trusted TLS-proxy address;
- a 1-30 second drain deadline whose minimum observation interval is shorter;
- at least one enabled game policy and an active scoped signing key.
Missing values produce an actionable startup error. The checked-in base file is
intentionally unsafe for Production so an accidental bare launch fails closed.
Signing keys support two external references:
- `env:NAME` reads 1-4096 bytes encoded as base64 from `NAME`;
- `file:/absolute/path` reads 1-4096 raw bytes from a non-symlink file.
Prefer a read-only container secret owned by the configured container identity.
For systemd, use a root-owned, `rendezvous`-group-owned `0440` file (or an
equivalent narrow ACL) so the non-root process can read but not replace it. A
signing key must contain at least 32 random bytes. Never put the key, publisher/operator
credential, or secret value in JSON, a command argument, an image layer, Compose
environment, logs, metrics, or source control. Configuration contains only the
reference and non-secret lifecycle metadata. A vault/KMS adapter can replace the
provider where local policy requires it.
Keep the host clock synchronized with authenticated NTP. Credential and key
windows use wall time; lease, timeout, drain, and rate-limit deadlines use a
monotonic clock. Alert on clock synchronization loss before rotating keys.
The checked-in Compose limits (one CPU and 512 MiB) are for its isolated smoke
profile, not a production capacity claim. The measured core-state candidate
uses 2 vCPU and 2 GiB with the same 128-PID/4096-descriptor ceilings; see
the [capacity and resilience gate](../operations/capacity-and-resilience.md).
Measure real traffic, then change resource limits and server budgets together.
Memory pressure or CPU throttling must not extend orchestrator termination past
`DrainDeadlineSeconds` plus five seconds.
## Graceful shutdown
SIGTERM and the authenticated operator drain both stop new registrations and
join attempts immediately. On process shutdown, HTTP and UDP remain available
long enough for existing join attempts to finish. The service exits as soon as
the minimum drain interval has elapsed and no attempts remain, or forcibly
clears all ephemeral state at the configured deadline. It then stops UDP and
HTTP listeners and exits. Configure Docker/systemd/Kubernetes termination grace
strictly longer than the service deadline; the examples use 40 seconds for a
30-second drain.
Never use SIGKILL for a normal rollout. After stopping, verify the process is
gone and neither `8080/tcp` nor `9050/udp` is bound before starting its
replacement on the same host. A crashed or force-killed process cannot drain;
clients recover through bounded retries and hosts re-register.
## systemd alternative
Publish the server for Linux, install the immutable output at `/opt/rendezvous`,
place production configuration beside the application read-only, place key
files below `/etc/rendezvous`, and install `deploy/systemd/rendezvous.service`:
```bash
dotnet publish src/FinalFactory.Rendezvous.Server \
--configuration Release --runtime linux-x64 --self-contained false \
--output publish/rendezvous
systemd-analyze verify deploy/systemd/rendezvous.service
sudo systemctl daemon-reload
sudo systemctl enable --now rendezvous.service
```
Create the dedicated `rendezvous` user without a login shell. Keep
`/opt/rendezvous` and `/etc/rendezvous` root-owned and non-writable by that user;
install each required key with `root:rendezvous` ownership and mode `0440`. The
unit applies the measured 2-vCPU/2-GiB core-state candidate profile plus the
same filesystem, privilege, network-family, and shutdown hardening as Compose.
## HTTP and UDP smoke
Build the diagnostic once, then exercise the actual published HTTP and UDP
paths. The test creates a public listing, sends authenticated presence and punch
traffic through UDP 9050, establishes peer-to-peer traffic, reports the outcome,
and deregisters cleanly:
```bash
dotnet build src/FinalFactory.Rendezvous.TestClient --configuration Release
./scripts/smoke-deployment.sh
```
For the local Compose profile, the script derives a ten-minute diagnostic
publisher credential from the ignored local key without printing either secret.
For production, do not copy the signing key to the smoke host. Instead inject a
short-lived, region-scoped credential through
`RENDEZVOUS_PUBLISHER_CREDENTIAL`, and set the external endpoints:
```bash
export RENDEZVOUS_PUBLISHER_CREDENTIAL='<short-lived deployment credential>'
export RENDEZVOUS_SMOKE_HTTP_URL='https://rendezvous.your-company.tld/'
export RENDEZVOUS_SMOKE_UDP_ENDPOINT='rendezvous-udp.your-company.tld:9050'
export RENDEZVOUS_SMOKE_GAME_ID='<credential game ID>'
export RENDEZVOUS_SMOKE_ENVIRONMENT_ID='<credential environment ID>'
export RENDEZVOUS_SMOKE_REGION='<credential region>'
export RENDEZVOUS_SMOKE_PROTOCOL_VERSION='<enabled protocol version>'
./scripts/smoke-deployment.sh
```
Those four scope values must match both the short-lived credential and an
enabled server policy. The defaults (`space-game`, `smoke`, `local`, protocol
`1`) are only for the checked-in local Compose profile.
The smoke fails unless both health endpoints and the complete authenticated UDP
mediation/direct-traffic flow succeed. It does not prove every consumer NAT;
run the topology harness and representative external-network tests as well.
## Restart, upgrade, rollback, and backup
Rendezvous has no durable runtime database. Restarting intentionally loses all
listings, observed endpoints, attempts, replay markers, and runtime-only
revocations. Hosts must treat registration as a renewable lease and re-register
after service recovery. Clients must re-browse and start a new bounded attempt.
Back up only reviewed configuration, policy, secret references, key material and
its custody/lifecycle records, deployment manifests, and image digest. Never
claim a backup contains live sessions or endpoints. Restore keys only through the
secret system, not into the image or repository.
For an upgrade:
1. Build and test the new pinned digest; validate configuration without starting
a second active instance.
2. Drain and stop the current process, verify both sockets are released, then
start the replacement on the same public endpoints.
3. Require live/readiness and HTTP+UDP smoke success; monitor host
re-registration, error rate, and direct-connect outcomes.
For rollback, repeat the same stop-before-start sequence with the previously
recorded image digest and compatible configuration/key set. Never run old and
new versions concurrently to avoid split ephemeral state. If a wire-incompatible
change ever becomes necessary, use a new API/protocol version rather than a
rolling two-version replica set.
@@ -0,0 +1,148 @@
{
"schemaVersion": 2,
"evidenceVersion": "v2",
"generatedAt": "2026-07-16T14:10:53.6981858+00:00",
"profile": "candidate",
"runtime": {
"framework": ".NET 10.0.9",
"operatingSystem": "CachyOS",
"kernel": "Unix 7.1.3.2",
"architecture": "X64",
"cpuModel": "AMD Ryzen 7 9800X3D 8-Core Processor",
"processorCount": 2,
"cpuAffinity": "0,1",
"cpuQuota": "not-enforced",
"memoryLimit": "not-enforced",
"garbageCollector": "workstation",
"commitSha": "cf14836d48b0b4aaa67f99433f4fba3585bcd2bb",
"treeState": "clean",
"command": "RENDEZVOUS_CAPACITY_PROFILE=candidate RENDEZVOUS_CAPACITY_CPUSET=0,1 ./scripts/run-capacity-gate.sh",
"imageDigest": "not-containerized",
"workloadSeed": "fixed-sequences-random-identifiers",
"capacityPhaseAverageCpuPercent": 56.37724115383554,
"peakWorkingSetBytes": 169705472,
"managedBytesAfterCleanup": 35615200
},
"targets": {
"visibleListings": 25000,
"activeJoinAttempts": 10000,
"coreControlOperationsPerSecond": 200,
"coreMediationOperationsPerSecond": 2000,
"coreControlP95Milliseconds": 200,
"coreMediationP95Milliseconds": 100,
"maximumAverageCpuPercent": 70,
"maximumWorkingSetBytes": 1610612736,
"soakCycles": 1000,
"soakDurationSeconds": 300
},
"measurements": [
{
"operation": "registration-and-presence",
"samples": 1000,
"p50Milliseconds": 0.003,
"p95Milliseconds": 0.0046,
"p99Milliseconds": 0.0054,
"operationsPerSecond": 282453.96000451926,
"minimumOperationsPerSecond": 200,
"budgetMilliseconds": 200,
"passed": true
},
{
"operation": "lease-renewal",
"samples": 1000,
"p50Milliseconds": 0.0004,
"p95Milliseconds": 0.0009,
"p99Milliseconds": 0.0021,
"operationsPerSecond": 968992.2480620155,
"minimumOperationsPerSecond": 200,
"budgetMilliseconds": 200,
"passed": true
},
{
"operation": "visible-session-browse",
"samples": 250,
"p50Milliseconds": 0.9046,
"p95Milliseconds": 3.3704,
"p99Milliseconds": 3.9471,
"operationsPerSecond": 695.5799787597697,
"minimumOperationsPerSecond": 200,
"budgetMilliseconds": 200,
"passed": true
},
{
"operation": "join-attempt-issuance",
"samples": 1000,
"p50Milliseconds": 0.0029,
"p95Milliseconds": 0.0045,
"p99Milliseconds": 0.0055,
"operationsPerSecond": 296428.042092782,
"minimumOperationsPerSecond": 200,
"budgetMilliseconds": 200,
"passed": true
},
{
"operation": "simultaneous-punch-pairing",
"samples": 1000,
"p50Milliseconds": 0.0043,
"p95Milliseconds": 0.0073,
"p99Milliseconds": 0.0115,
"operationsPerSecond": 109212.03516627532,
"minimumOperationsPerSecond": 2000,
"budgetMilliseconds": 100,
"passed": true
},
{
"operation": "principal-revocation",
"samples": 50,
"p50Milliseconds": 0.518,
"p95Milliseconds": 0.7049,
"p99Milliseconds": 11.8557,
"operationsPerSecond": 1320.1773262184577,
"minimumOperationsPerSecond": 50,
"budgetMilliseconds": 200,
"passed": true
},
{
"operation": "telemetry-recording",
"samples": 1000,
"p50Milliseconds": 0.0001,
"p95Milliseconds": 0.0001,
"p99Milliseconds": 0.0001,
"operationsPerSecond": 1438641.9220256077,
"minimumOperationsPerSecond": 10000,
"budgetMilliseconds": 1,
"passed": true
},
{
"operation": "coincident-listing-attempt-expiry",
"samples": 1,
"p50Milliseconds": 29.6882,
"p95Milliseconds": 29.6882,
"p99Milliseconds": 29.6882,
"operationsPerSecond": 33.682962483916384,
"minimumOperationsPerSecond": 0,
"budgetMilliseconds": 200,
"passed": true
}
],
"state": {
"peakListings": 25000,
"peakAttempts": 10000,
"peakReplayMarkers": 0,
"finalListings": 0,
"finalAttempts": 0,
"finalReplayMarkers": 0,
"expiryChurn": 94906,
"maintenanceSweeps": 36307,
"soakCyclesCompleted": 75126848,
"soakDurationSeconds": 300.0000015,
"soakPeakScheduledExpiryEntries": 7,
"soakManagedGrowthBytes": -257288,
"soakHandleGrowth": 2,
"restartStartedEmpty": true,
"overloadWasTyped": true,
"recoverySucceeded": true
},
"failures": [],
"passed": true
}
+97
View File
@@ -0,0 +1,97 @@
# Diagnostic TestClient integration guide
Tracking: #25
`FinalFactory.Rendezvous.TestClient` is the smallest supported public-SDK consumer.
It exists for integration development, CI smoke checks, deployment verification,
and operator diagnosis. It is intentionally not a production game client, game
server, matchmaking UI, or relay.
The automated scenario matrix, privileged Linux namespace run, and topology
limitations are documented in the [deterministic topology harness](topology-harness.md).
## Prerequisites
Start a configured Rendezvous service and note both its HTTP base URL and UDP
mediator endpoint. The host needs a tenant-scoped publisher credential from the
deployment secret boundary. Put it in an environment variable and pass only that
variable's name when the default is unsuitable:
```bash
export RENDEZVOUS_PUBLISHER_CREDENTIAL='<deployment-supplied value>'
```
Never put the credential in a command argument, URL, checked-in configuration,
shell trace, or captured test fixture. The development server's signing material
is process-ephemeral; credentials from a prior development process are invalid.
## Manual three-terminal flow
Start the host:
```bash
dotnet run --project src/FinalFactory.Rendezvous.TestClient -- \
host --service http://127.0.0.1:5000/ --mediator 127.0.0.1:9050 \
--game space-game --environment development --region local --protocol 1
```
Browse from another terminal:
```bash
dotnet run --project src/FinalFactory.Rendezvous.TestClient -- \
browse --service http://127.0.0.1:5000/ \
--game space-game --environment development --region local --protocol 1
```
Join from a third terminal. Omit `--listing` for an interactive choice:
```bash
dotnet run --project src/FinalFactory.Rendezvous.TestClient -- \
join --service http://127.0.0.1:5000/ --mediator 127.0.0.1:9050 \
--game space-game --environment development --region local --protocol 1 \
--listing 00000000-0000-0000-0000-000000000000
```
Replace the sample UUID with the public listing ID printed by host or browse.
Host and join each create one caller-owned LiteNetLib manager. That same socket
sends presence/punch traffic, establishes the authenticated direct connection,
and carries the ping/echo/ack/completion payload. The final completion confirms
that the host received the reliable acknowledgement; none of this traffic passes through the HTTP
service or UDP mediator.
## CI and deployment smoke flow
Use `--script --json`, set `--listing` when deterministic selection matters, and
check the documented process exit code. `--timeout-seconds` bounds each startup,
traversal, or direct-traffic stage; a script host also uses it as its total runtime
unless `--run-seconds` is explicit. A host can add `--exit-after-echo` so it
terminates after the joining peer acknowledges direct traffic and receives the
host's completion confirmation. Every wait is
bounded by coordinator state and `--timeout-seconds`; no orchestration should use
an unbounded sleep.
The normal test suite contains a real process gate that starts the built Server,
host TestClient, and join TestClient, waits for readiness and versioned events,
and verifies direct traffic, cleanup, JSON shape, and secret canaries. Process
trees are force-terminated in the test cleanup path if normal shutdown fails.
Useful success events are:
- `host.registered`, `host.ready`, `host.direct-traffic`, and `host.deregistered`;
- `browse.completed` and `browse.session`; and
- `join.connected`, `join.direct-traffic`, and `join.outcome-report`.
Failure events preserve stable typed phases and outcomes. When a terminal outcome
contains a configured dedicated endpoint, `join.fallback` reports `available`
with endpoint type `dedicated`; no raw address is printed and no fallback is
started implicitly.
## What the proof does and does not establish
The deterministic loopback test proves the complete service/host/client protocol,
ticket admission, and peer-to-peer payload path. Loopback is not evidence that all
consumer routers, carrier-grade NATs, symmetric NATs, firewalls, VPNs, IPv6 paths,
or platform policies permit hole punching. Same-LAN, separated observed endpoints,
network namespaces/containers, mediator restart, and adverse topology coverage
belong to the topology harness tracked by #14. Production rollout still requires
tests from representative networks and a game-owned fallback policy.
+91
View File
@@ -0,0 +1,91 @@
# Deterministic topology harness
Issue #14 is verified at three layers. The layers are deliberately separate so
the always-on gate remains deterministic while privileged CI workers can add a
stronger operating-system topology without overstating what local emulation
proves about the public Internet.
## Always-on public-process gate
`TestClientProcessIntegrationTests` launches the built server and the same
`FinalFactory.Rendezvous.TestClient` executable shipped to operators. Every
child process uses `--script --json`, dynamic HTTP and UDP ports, bounded
state-driven waits, and enforced process-tree cleanup.
The suite proves:
| Scenario | Required observation |
| --- | --- |
| Three-party happy path | register, presence-ready, browse, authorize, punch, authenticated LiteNetLib connection, direct ping/echo/ack/completion traffic, outcome report, disconnect, deregister |
| Same-LAN candidate | the connected peer is reported as `loopback` or `private`, never inferred merely from an introduction callback |
| Empty and missing selection | browse exits `11`; exact missing lookup exits `10` |
| Wrong tenant/protocol | no listing is returned for an incompatible protocol; exact joins with either mismatch fail before `join.punch` |
| Traversal timeout | an unreachable mediator produces typed `PunchTimedOut`, exits `12`, advertises the configured dedicated fallback, and never connects to it |
| Caller cancellation | POSIX `SIGINT` exits `130`, deregisters the listing, and removes it from public lookup |
| Abrupt host loss | the listing disappears after the presence window and before its lease expires; public exact lookup intentionally reports `NotFound` |
| Bounded host without a peer | exits `13` and still deregisters |
Captured output is parsed as the stable JSON v1 event schema. Publisher
credentials and signing-key material are checked against all captured output.
The direct traffic payload is handled only by the caller-owned host and client
LiteNetLib managers; the HTTP service and mediator do not implement or observe
the echo protocol.
Run the always-on scenarios with:
```bash
dotnet test Rendezvous.slnx --configuration Release --no-build \
--filter FullyQualifiedName~TestClientProcessIntegrationTests
```
## Deterministic protocol and adverse-state gate
The following real service-boundary tests cover conditions that a public CLI
cannot safely manufacture by accepting raw capabilities or tickets:
| Scenario | Test evidence |
| --- | --- |
| Same-NAT private candidates | `NatMediationProcessorTests.MatchedPeersReceiveOneIntroductionAndSameNatPrivateCandidates` |
| Separate observed endpoints | `NatMediationProcessorTests.DifferentNatsAndInvalidLocalClaimsExposeOnlyObservedPublicEndpoints` |
| One-time introduction and replay | `InMemoryEphemeralRendezvousStoreTests.AttemptCapabilitiesAndIntroductionAreOneTime` |
| Direct ticket replay | `RendezvousCoordinatorIntegrationTests.CallerOwnedManagersCompleteAuthenticatedDirectConnectionAndRejectTicketReplay` |
| Wrong tenant/protocol and stale presence | `InMemoryEphemeralRendezvousStoreTests.JoinRequiresExactScopeProtocolAndFreshHostPresence` |
| Cancellation and late callbacks | `RendezvousCoordinatorBehaviorTests.CancellationCompletesExactlyOnceAndLateCallbacksCannotReopenTheAttempt` |
| Mediator restart | both cases of `UdpMediatorServiceTests.NativeLiteNetLibRequestsIntroduceTheAuthorizedPair`; the restarted case rebinds the same UDP port and completes a native LiteNetLib introduction |
These tests use fake monotonic clocks or state predicates where expiry and race
ordering matter. They do not use fixed sleeps as proof of state.
## Privileged Linux namespace gate
When a Linux CI worker can create network namespaces, the workflow sets
`RENDEZVOUS_RUN_NETNS_TESTS=1` and reruns
`PrivilegedLinuxNatNamespacesCompleteDirectTrafficAcrossSeparateObservedEndpoints`.
The test creates a temporary WAN bridge, an isolated service namespace, two NAT
router namespaces, and isolated host/client LAN namespaces. Each NAT has its own
inside subnet and WAN address. Linux forwarding plus per-router MASQUERADE rules
force the service to observe separate translated endpoints; the public TestClient
processes must then complete authenticated direct traffic through those mappings
using the public candidate. Namespaces, rules, veth pairs, bridge, processes, and
sockets are removed in bounded async-disposal paths. A cleanup failure fails the
test.
If `ip netns add`/`iptables` is unavailable or the worker lacks `CAP_NET_ADMIN`,
CI records the limitation and keeps the always-on loopback suite as the required gate.
To request the privileged run explicitly:
```bash
RENDEZVOUS_RUN_NETNS_TESTS=1 dotnet test Rendezvous.slnx \
--configuration Release --no-build \
--filter FullyQualifiedName~PrivilegedLinuxNatNamespacesCompleteDirectTrafficAcrossSeparateObservedEndpoints
```
## What this does not prove
Loopback, MASQUERADE, and namespace routing cannot reproduce every consumer router,
carrier-grade NAT, firewall, IPv6 transition mechanism, symmetric NAT mapping,
or real-world packet-loss pattern. The separate-observed-endpoint processor
test proves that untrusted private claims are excluded and public candidates are
selected; it is not presented as universal Internet traversal proof. Real
network canaries and measured production readiness remain the scope of issue
#23.
+189
View File
@@ -0,0 +1,189 @@
# Capacity, resilience, and availability gate
Tracking: #18
This gate turns the v1 budgets in ADR 0003 into a repeatable release decision.
It does not turn Rendezvous into a horizontally scalable service: v1 remains one
active process with bounded in-memory state. A second process may be a cold
standby, but it must not accept traffic until the first process has stopped and
released the public HTTP and UDP endpoints.
## Launch envelope and approved core-state profile
The approved core-state profile is one Linux process limited to 2 vCPU and
2 GiB RAM. Public HTTP/UDP numbers are launch objectives that require the #23
real-network canary before they become a supported service claim:
| Dimension | Value | Evidence status |
| --- | --- | --- |
| Visible listings | 25,000 | Enforced and measured here |
| Active join attempts | 10,000 | Enforced and measured here |
| Core control path | 200 operations/second; p95 at most 200 ms | Measured here |
| Core mediation path | 2,000 pairings/second; p95 at most 100 ms | Measured here |
| Sustained HTTP demand | 200 requests/second | #23 launch objective; not yet a supported claim |
| Sustained UDP demand | 2,000 datagrams/second | #23 launch objective; not yet a supported claim |
| Public HTTP/UDP latency | p95 at most 200 ms / 100 ms | #23 launch objective; not yet a supported claim |
| Capacity-phase average CPU / peak working memory | below 70% / below 1.5 GiB | Measured for the core candidate |
| Valid in-profile monthly availability | 99.5%, excluding announced maintenance | Operational objective |
| Process-ready RTO / host-visible recovery | 15 seconds / 90 seconds | 15 seconds automated; 90-second deployment drill required |
The proposed public-network mix is 20% registration/update, 30% lease-critical
renew/delete, 30% browse, and 20% join authorization for HTTP. The UDP mix is
60% authenticated host-presence refresh, 30% attempt contributions, and 10%
invalid or duplicate traffic that must be dropped early. A deployment may use a
lower per-game profile, but must not claim a higher one without new versioned
evidence.
The capacity harness fills the complete state ceilings, then measures
registration plus presence, renewal, a 100-item compatible browse, join
issuance, simultaneous two-peer pairing, principal revocation, and telemetry.
It applies 200/100 ms guardrails and minimum 200 control / 2,000 mediation
operations per second to the core hot path. Those measurements deliberately
exclude Kestrel, LiteNetLib, TLS, JSON, socket scheduling, and the documented
mixed traffic shape. The #23 real-network canary must exercise those layers,
rate-shape the mix, record errors and shedding, and meet the public objectives
before launch; a core result is not a public-network latency or throughput claim.
## Reproduce the evidence
Every push runs the quick profile and the selected fault matrix:
```bash
./scripts/run-capacity-gate.sh
```
Run the production candidate on an otherwise idle Linux host and restrict the
runtime to two logical CPUs. The default candidate includes a five-minute,
high-intensity expiry soak; use 3,600 seconds for a release-candidate endurance
run:
```bash
export RENDEZVOUS_CAPACITY_PROFILE=candidate
export RENDEZVOUS_CAPACITY_CPUSET=0,1
export RENDEZVOUS_CAPACITY_OUTPUT="$PWD/artifacts/capacity/candidate.json"
./scripts/run-capacity-gate.sh
# Release-candidate endurance override:
dotnet run --project tests/FinalFactory.Rendezvous.Capacity \
--configuration Release --no-build -- \
--profile candidate --soak-seconds 3600 \
--output artifacts/capacity/candidate-endurance.json
```
The machine must have at least 2 GiB available to the process. For formal
deployment evidence, run inside the same cgroup/container shape as production.
The v2 JSON embeds the commit and tree state, command, image context, CPU model,
kernel, affinity, cgroup quota/limit, collector mode, and workload seed. Supply
`RENDEZVOUS_EVIDENCE_IMAGE_DIGEST` when running a release image. Do not compare
results collected under a debugger,
concurrent build, thermal throttling, or oversubscribed CI host.
The checked-in baseline is
[`candidate-2cpu.json`](../evidence/capacity/v2/candidate-2cpu.json). It was
produced on .NET 10.0.9/Linux x64 with CPU affinity restricted to two logical
CPUs. It filled 25,000 listings and 10,000 attempts, peaked at about 162 MiB,
and cleared all active/retained state. The five-minute baseline supersedes any
earlier local probe when its timestamp and target duration differ.
## Soak and bounded-state interpretation
Each soak cycle creates a listing, repeatedly renews its lease and refreshes
presence, creates a join attempt, replay marker, and retained outcome, checks
that scheduled expiry entries remain proportional to live keys, then advances
the injected monotonic clock beyond all
deadlines, and verifies that listings, attempts, replay, idempotency, and outcome
state return to zero. The candidate also measures managed-memory and process
handle deltas after full collection. Failure is any retained state, more than
64 MiB retained managed memory, more than eight retained handles, a working set
above 1.5 GiB, an untyped capacity result, or failure to admit work after expiry.
This accelerated soak intentionally executes far more state lifecycle/cleanup
events than wall-clock traffic would permit. It catches stale deadline-queue
entries, cache growth, replay/idempotency retention, and cleanup cost. Because
it does not open Kestrel/LiteNetLib connections, its process-handle delta is only
a harness guard and is not evidence of transport stability by itself. The
selected production-process gate adds a ten-second real HTTP/UDP transport soak,
samples child-process handles and RSS, asserts bounded growth, then verifies a
clean SIGTERM and socket release. #23 must extend that into the full rate-shaped
multi-client canary while sampling queues, managed memory, and state
cardinalities. A one-hour core override remains required before tagging a
production release.
## Fault and recovery matrix
`run-capacity-gate.sh` runs these deterministic production paths before the
numeric profile:
| Fault | Required result |
| --- | --- |
| HTTP/UDP overload and tracker exhaustion | Typed HTTP `429`/`CapacityExceeded`, silent UDP drop, bounded tracker keys, recovery after the window |
| Optional traffic saturation | Lease-critical renew/update/delete capacity remains available |
| Store/dependency unavailable | Readiness fails; new authorization returns typed `ServiceUnavailable`; liveness remains independent |
| Graceful drain/SIGTERM | New work returns `Draining`; existing pairing may finish; process exits 0 and releases TCP/UDP before the deadline |
| Hard restart | In-flight state is lost; SDK reports typed `ServiceUnavailable`; a host re-registers, rebinds presence, and becomes the only browser-visible replacement |
| UDP listener bind/restart | Readiness stays false without the required listener; rebinding the advertised port restores native LiteNetLib pairing |
| Wall-clock jump/skew | Monotonic lease/attempt authority is neither shortened nor extended; credential skew remains capped at 30 seconds |
| Signing-secret rotation | New key signs, overlap verifies, retired/revoked key rejects, missing material fails startup |
| Principal revocation | Listing, presence, attempts, and outcome paths are removed atomically within the latency budget |
No external database exists in v1, so “dependency/store failure” means the
process-local atomic store is marked unavailable or a required listener/key is
unready. The service fails closed rather than pretending a degraded writable
mode exists.
## Bandwidth and amplification
- Accepted application datagrams are at most 1,200 bytes.
- Malformed, oversized, unauthenticated, stale, replayed, wrong-role, and
rate-limited traffic receives zero response bytes.
- A completing authenticated contribution produces at most one introduction to
each observed peer, and the combined response is at most 2.0 times that
contribution's bytes.
- The frozen-envelope and native LiteNetLib socket tests measure this on the real
UDP listener; the hostile corpus and allocation gate exercise 10,000+ inputs
without input-sized logs, tasks, or queues.
Bandwidth planning must therefore reserve ingress for the configured 2,000
datagrams/second plus edge overhead and egress for a worst-case verified 2.0
amplification. Actual successful pairs normally use two contributions and two
introductions; normal gameplay leaves Rendezvous entirely.
## Availability decision
Single-active remains the v1 topology. The measured core profile proves bounded
state and substantial core-path headroom, while public launch capacity remains
conditional on #23. The service has a bounded stop-before-start restart path.
Its failure domain is deliberately
one process/node/public UDP endpoint: node, kernel, host network, DNS/TLS edge,
secret configuration, or operator error can remove all readiness until the cold
replacement owns the same source-preserving endpoint.
The 99.5% objective permits about 216 minutes of unannounced downtime in a
30-day month. Operations must target process readiness within 15 seconds and
host-visible re-registration within 90 seconds, page when no ready instance
exists, and include detection plus recovery in the monthly budget. The current
in-process test validates typed downtime, same-port HTTP restart, fresh
registration, presence rebinding, and browser visibility in under five seconds;
the production-process test separately validates graceful termination, TCP/UDP
release, replacement startup on the same endpoints, UDP readiness, and the
15-second process-ready RTO. Cold-standby activation policy and the 90-second
operator-to-host recovery objective still require a deployment drill before
release. Rollout and rollback use the
deployment runbook's drain, stop, socket-release, start, smoke sequence; never
overlap old and new active processes.
Bring shared TTL/CAS state and deterministic mediator routing forward before
enabling two active instances if any of these occurs:
- one node cannot sustain 150% of the measured 30-day peak while meeting SLOs;
- CPU stays above 70%, memory above 75%, attempt depth above 70%, or limiter
drops/latency remain elevated after abusive traffic is excluded;
- the availability target rises above 99.5% or planned maintenance must preserve
listings; or
- one region requires multiple simultaneously active mediator endpoints.
Rendezvous makes no multi-instance claim today, so a two-node atomic-pairing
test is intentionally not applicable. It becomes a hard release gate with the
shared-state/routing implementation; until then `SingleActiveInstance=false`
fails production startup. Multi-region and relay remain separate evidence-driven
decisions.
@@ -0,0 +1,144 @@
# Observability and operator runbook
This runbook defines the production signals and privileged controls for the
Rendezvous service. The service emits `System.Diagnostics.Metrics` instruments
from the `FinalFactory.Rendezvous` meter and distributed-tracing activities from
`FinalFactory.Rendezvous.Server`. Connect those sources to the deployment's
OpenTelemetry or equivalent collector. Do not add identifiers to metric labels.
## Health and readiness
- `GET /health/live` proves that the HTTP process can answer. It deliberately
remains independent of provisioning, the state store, drain state, and optional
listeners so an orchestrator does not restart a recoverable dependency failure.
- `GET /health/ready` returns success only after the HTTP path is answering, the
required IPv4 UDP socket is bound, any configured IPv6 UDP socket is bound,
provisioning loaded successfully, the store is available, and drain has not
started. A failed check returns `503` and removes the instance from new work.
- A graceful drain immediately makes readiness fail while liveness remains healthy.
Existing work may complete until the bounded store drain deadline.
## Metrics and traces
| Instrument | Purpose | Bounded dimensions |
| --- | --- | --- |
| `rendezvous.http.requests` / `rendezvous.http.duration` | HTTP volume and latency | operation, status code |
| `rendezvous.udp.results` / `rendezvous.udp.duration` | UDP mediation volume and processing latency | frozen/litenet operation, result |
| `rendezvous.limiter.drops` | Requests shed by admission controls | transport, fixed partition class |
| `rendezvous.operator.authentication` | Accepted, forbidden, and rejected operator authentication | result |
| `rendezvous.audit.events` | Privileged action outcomes | fixed action, result |
| `rendezvous.connection.outcomes` | Client-reported direct-connect outcomes | normalized outcome, elapsed bucket |
| `rendezvous.pairing.latency` | Time from attempt creation to successful peer introduction | none |
| `rendezvous.queue.depth` | Active join-attempt queue depth | none |
| `rendezvous.store.active_listings` / `active_leases` / `active_attempts` / `replay_markers` | Current ephemeral load | none |
| `rendezvous.store.expiry_churn` | Cumulative natural expiry activity | none |
| `rendezvous.store.available` | Store health (`1` available, `0` unavailable) | none |
HTTP responses include `X-Rendezvous-Correlation-ID`. It is a generated trace ID
or random value, never a caller-supplied session or player identifier. UDP and
HTTP activities contain operation-level data only. Logs and traces must not add
tokens, capabilities, session/listing IDs, player subjects, metadata, raw IP
addresses, or endpoint values.
Recommended dashboard panels are request rate and p50/p95/p99 latency by fixed
operation, UDP result ratio, direct connection success ratio, pairing latency,
active listings/attempts, expiry churn, limiter drops, store availability,
operator authentication results, audit action results, and signing-key windows.
## Alerts
Tune thresholds from the normal production baseline, then keep these conditions
as distinct actionable alerts:
- **Signing key expiry:** page when any required signing key has less than seven
days before `signUntil`; escalate at 24 hours. Confirm a replacement is signing
and the previous key remains verify-only for the maximum credential lifetime.
- **Authentication spike:** warn when rejected or forbidden operator authentication
exceeds five attempts in five minutes. Treat unexpected publisher-authentication
growth as a possible credential or integration incident.
- **Direct success regression:** warn when the connected outcome ratio falls more
than 20% below its seven-day same-region baseline for 15 minutes, with a minimum
sample floor. Break down only by bounded outcome and time bucket.
- **Saturation:** warn when queue depth remains above 70% of the configured attempt
limit, limiter drops are sustained, or p95 latency exceeds the service objective;
page at 90% or when lease-critical traffic is shed.
- **Store degradation:** page immediately when `rendezvous.store.available` is zero
or readiness fails for the store. Rising expiry churn without corresponding new
work is a warning for stalled clients or clock/configuration mistakes.
- **Listener/config readiness:** page when no ready instances remain. Investigate
UDP bind failures, a configured-but-unbound IPv6 listener, provisioning errors,
and unintended drain state separately.
## Operator authentication and controls
Operator credentials use a signing key configured with `CredentialKinds:
["Operator"]`. Operator keys cannot be scoped to a game/environment or used for
publisher credentials. Mint short-lived operator credentials through the trusted
provisioning process, outside the public Rendezvous HTTP service, and grant only
the required permission. Never place credentials in command history, URLs, logs,
or support tickets.
The application also enforces a default-deny source boundary. Configure at most
32 exact operator source IPs in
`Rendezvous:AbuseProtection:OperatorAllowedAddresses`; an empty list disables all
operator HTTP access. Development permits loopback only. Production must place
`/v1/operator/*` behind a private management listener or reverse-proxy ACL, list
only the resulting trusted management source addresses, and block that path on
the public edge. If forwarded headers are enabled, keep the existing exact-proxy,
single-hop trust policy and allowlist the post-forwarding operator source. Verify
from both an allowed management host and a denied public host before deployment.
Denied sources are charged to the bounded general HTTP partition before credential
or request-body processing, then receive `404`; sustained denied traffic receives
the same typed `429` overload response as other public traffic.
Operator traffic has a dedicated, bounded rate/concurrency partition and critical
tracker-key reserve. Public browse/join saturation therefore cannot consume the
operator control budget, while compromised management sources remain rate-limited.
The OpenAPI document defines the separate `OperatorBearer` scheme. All endpoints
are under `/v1/operator`:
| Endpoint | Permission | Confirmation |
| --- | --- | --- |
| `GET /status` | `ReadPolicy` | none; returns aggregates, tenant status, safe key status, and audit counts |
| `POST /listings/revoke` | `RevokePublisher` | repeat the exact listing ID in `confirmListingId` |
| `POST /principals/revoke` | `RevokePublisher` | repeat the exact subject and choose a 1-600 second revocation lifetime |
| `POST /keys/revoke` | `RotateKeys` | repeat the exact key ID; runtime revocation is immediate |
| `POST /drain` | `ManagePolicy` | send the exact value `DRAIN` |
Publisher credentials are rejected on this surface even if their subject resembles
an operator. Destructive responses do not echo identifiers. The status response
does not expose player identities, raw endpoints, session metadata, capabilities,
or tokens. Every authenticated operator action, rejected confirmation, and
permission denial is audited with actor and target fingerprints.
Key revocation is process-local in the current single-instance store. Apply the
same revocation to every instance, then replace configuration before restarting;
a restart reconstructs the configured key ring. Principal revocation is bounded
to ten minutes and removes that principal's active listings and attempts. A
repeat action may extend an active revocation but never shortens it; wait for its
original deadline rather than treating a shorter repeat as an un-revoke. Use
listing revocation for one targeted session and drain before planned shutdown.
## Audit retention and incident handling
The in-process audit trail defaults to 10,000 entries and 30 days. It evicts the
oldest record at capacity and purges expired records on the next write. Configure
`Rendezvous:Audit:MaxEntries` and `RetentionDays` within their validated bounds.
Export the structured `AuditTrail` log events through the deployment's protected
logging pipeline when durable retention is required; the in-memory trail is not a
durable compliance archive. Those events include only timestamps, fixed action
fields, correlation IDs, and actor/target fingerprints.
Audit records retain timestamp, fixed action/result, target kind, correlation ID,
and 96-bit SHA-256 fingerprints of actor and target. Routine logs contain only the
fixed action/result/target kind and correlation ID. Restrict audit access to the
operator role, retain aggregates only as long as operationally necessary, and
delete raw exported audit data according to the 30-day policy unless an incident
hold is approved.
During an incident: confirm readiness and store health; capture aggregate graphs
and correlation IDs; revoke the narrowest listing, principal, or key; drain only
when isolation is required; record the action in the incident timeline; and verify
that direct success, limiter drops, and authentication rates return to baseline.
Do not copy player data, endpoints, or credentials into the incident record.
+150
View File
@@ -0,0 +1,150 @@
# Releases and compatibility
Tracking: #19
Rendezvous releases are immutable, reproducible, and promoted only after the
same candidate has passed package, consumer, server, container, and staging
checks. A release consists of matching Client and Contracts NuGet packages, a
framework-dependent Linux server archive, a versioned linux/amd64 OCI image,
package/runtime and container SPDX inventories, checksums, provenance, release
notes, and signatures. No workflow publishes a `latest` tag.
## Version dimensions
The central values in `eng/Versions.props` are the authority. Client,
Contracts, and Server use SemVer. HTTP, UDP mediation, and connection-ticket
formats advance independently so a wire change cannot hide inside a package
patch release. The current machine-readable matrix is
[`compatibility.json`](compatibility.json); the authenticated operator status
endpoint exposes the server's supported window at runtime.
| Surface | Current | Compatibility rule |
| --- | ---: | --- |
| Client and Contracts | 1.0.0 | Matching exact versions; source/API breaks require a package major bump. |
| Server | 1.0.0 | Accepts Client 1.0.0 through compatible 1.x releases. |
| HTTP contract | 1 | Frozen OpenAPI, JSON vectors, and public API snapshot. |
| UDP mediation | 1 | Frozen codec vectors; incompatible bytes require UDP v2. |
| Connection ticket | 1 | A format change requires a new accepted ticket version and migration window. |
| LiteNetLib | 2.1.4 | Exact dependency; LiteNetLib 1.x is rejected by package and consumer gates. |
| Gameplay protocol | Per tenant | Exact match; Rendezvous does not translate gameplay protocols. |
`scripts/check-compatibility.sh` compares protected snapshots against the base
revision. A changed public API snapshot requires a package major increase; an
HTTP or UDP golden surface requires the corresponding contract increase. The
normal tests also compare implementation output with the current versioned
snapshots. For a deliberate break, add a new versioned contract directory and
documentation instead of replacing the prior version's evidence.
## Candidate build
From a clean tagged checkout:
```bash
./scripts/check-release-tag.sh v1.0.0
./scripts/build-release.sh 1.0.0
./scripts/verify-release.sh 1.0.0
```
The tag build runs inside the digest-pinned `release-builder` Docker stage,
which combines the pinned SDK with a pinned Python runtime. It uses the locked
dependency graph, enforces NuGet
advisories and approved licenses, runs formatting/build/tests, regenerates the
OpenAPI drift check, and packs twice after a clean rebuild. NuGet's random OPC
relationship identifiers are canonicalized before comparison; both `.nupkg`
and `.snupkg` outputs must then be byte-identical. Package metadata identifies the exact
repository commit, portable PDBs carry SourceLink data, and the Linux archive,
runtime SBOM timestamp, and checksum ordering are deterministic. Buildx and
BuildKit are also pinned for the linux/amd64 OCI build. Provenance records each
artifact-producing tool version; an out-of-band rebuild must use the pinned
builder and recorded versions rather than treating the runner label as a
reproducibility guarantee.
All project-authored artifact normalization and checksum updates run inside the
same pinned builder; host tools only orchestrate or verify. The separately
pinned Trivy and Cosign tools produce the container inventory and signatures.
The tag workflow also performs two no-cache image builds with the commit time
and revision fixed, disables unsigned builder-generated attestations, and
requires identical OCI image IDs before signing the project provenance.
The local gate builds net8.0 SpaceGame- and Unscouted-shaped API fixtures using
only the candidate feed plus NuGet.org; the Unscouted fixture also carries its
real direct LiteNetLib 2.1.4 pin. The tag gate separately checks out the exact
SpaceGame and Unscouted revisions in `eng/consumer-revisions.json`, injects
exact candidate references without modifying those repositories, and restores
their real game/network projects. Both paths must resolve the matching Client
and Contracts version and LiteNetLib 2.1.4. Updating a consumer revision is a
reviewed compatibility change, not a floating-main check.
## Promotion and publication
Pushing the matching `vMAJOR.MINOR.PATCH` tag starts the tag-only release
workflow. Before any external write it:
1. builds and verifies the artifact set;
2. builds the exact versioned container candidate;
3. starts that image with production hardening and a temporary staging key;
4. completes HTTP health, registration, browse, authenticated UDP mediation,
and direct traffic;
5. rejects all high or critical container findings and emits a container SPDX
inventory;
6. finalizes and verifies checksums over the publish-ready artifact set; and
7. confirms the two NuGet versions, container version, and Gitea release do not
already exist.
Publication has no skip-duplicate behavior. Gitea's immutable package versions,
the workflow concurrency lock, and the preflight make a successful tag a
single publication event. The workflow pushes symbols, publishes only the
versioned container tag, records its `sha256` digest in both a digest file and
provenance, regenerates the checksum manifest so that digest and public key are
covered, signs the checksum manifest and image, attaches signed provenance,
verifies the complete published-set schema and all signatures, and creates the
Gitea release with exactly those artifacts. The detached checksum signature
bundle is the sole envelope excluded from its own signed manifest.
The loaded image ID is captured immediately after the byte-reproducible build;
publication refuses to push if staging or another process retagged that local
name to different bytes.
The protected `production` environment requires these secrets:
- `RELEASE_TOKEN`: a dedicated Gitea token limited to this repository and the
HeiKyu package registry, with repository and package write access;
- `RELEASE_USERNAME`: the dedicated Gitea service-account name that owns the
release token;
- `COSIGN_PRIVATE_KEY`: the encrypted Cosign release private key; and
- `COSIGN_PASSWORD`: its password, stored separately.
No development signing key, registry credential, or deployable configuration
is stored in source or packages. Keep the Cosign public key with operational
records. Rotate the release key between releases: retain the old public key for
historical verification, install the new encrypted private key and password as
one reviewed change, verify a signed non-release blob, and only then retire the
old secret. Suspected compromise requires token/key revocation and a new
version; never overwrite or delete evidence to reuse a released version.
## Release notes and migration
Every dated `CHANGELOG.md` entry must contain Compatibility, Security and
configuration, and Migration sections. Before tagging, state the supported
Client/server window, all HTTP/UDP/ticket changes, security fixes, required
configuration, and operator/consumer migration steps.
For a protocol migration, first make the server read both old and new versions
within an explicit bounded window, publish a Client that writes the new version,
verify adoption through bounded telemetry, then remove the old reader only in a
subsequent breaking release. Never silently reinterpret old bytes. Consumers
pin both Rendezvous packages to one exact version and choose gameplay protocol
compatibility per tenant.
## Rollback and interrupted publication
Runtime rollback means redeploying the previous known-good image by digest and
its matching configuration; it does not move a tag. Packages and release
records remain available so already restored clients stay reproducible. If a
new release is faulty, revoke affected publisher or signing keys when relevant,
mark the release notes as withdrawn, and publish the fix under a new SemVer.
The registries cannot provide a transaction spanning NuGet, OCI, signatures,
and release attachments. If publication stops after its first external write,
the preflight intentionally prevents an automatic rerun. An operator must
inventory every destination, preserve logs and hashes, complete or withdraw the
partial version under change control, and then issue a new version. This avoids
turning a partial failure into an untraceable overwrite.
+27
View File
@@ -0,0 +1,27 @@
{
"schemaVersion": 1,
"release": "1.0.0",
"server": {
"minimumClientVersion": "1.0.0",
"maximumClientMajorVersion": 1
},
"packages": {
"FinalFactory.Rendezvous.Client": "1.0.0",
"FinalFactory.Rendezvous.Contracts": "1.0.0"
},
"contracts": {
"http": [1],
"udp": [1],
"connectionTicket": [1],
"gameplay": "exact-per-tenant"
},
"transport": {
"package": "LiteNetLib",
"version": "2.1.4",
"major": 2
},
"consumers": {
"SpaceGame": "net8.0",
"Unscouted": "net8.0"
}
}
+103
View File
@@ -0,0 +1,103 @@
# Hostile-input and overload protection
Tracking: #15
Rendezvous treats every public HTTP request and UDP datagram as hostile. The
server applies bounded fixed-window request budgets and concurrency ceilings in
two stages so malformed input is discarded before expensive work while valid
traffic is also isolated by its authenticated scope.
## Enforcement order
1. Kestrel and the HTTP abuse middleware cap request bodies at 16 KiB. A known
oversized body receives a typed `413` response before endpoint dispatch.
2. Every HTTP request consumes global, source-prefix, and operation budgets and
acquires the corresponding concurrency leases. IPv4 sources share a `/24`
budget and IPv6 sources share a `/56` budget; raw addresses are not retained.
Non-lease operations also consume a smaller optional-work budget, leaving a
configured global and source-prefix reserve for renew, update, and delete
operations during shedding.
Health probes use their own source-prefix budget so public API overload cannot
make a healthy instance fail its orchestrator probes, while health traffic is
still bounded.
Operator endpoints likewise use a separate bounded rate/concurrency partition
backed by the critical tracker reserve. They first require an exact source IP
from the default-deny `OperatorAllowedAddresses` policy, so public traffic
cannot spend the incident-response budget.
3. Once an endpoint has safely derived identities, it also acquires applicable
tenant, principal or capability, and listing/attempt budgets. Secret
capabilities are represented only by bounded SHA-256 fingerprints.
4. Every UDP envelope consumes global, source-prefix, and wire-operation
budgets before decoding. A structurally and cryptographically valid request
then consumes capability, role, and mediation-handle budgets before state
mutation or introduction.
5. HTTP overload returns the stable `RateLimited` error, status `429`, and a
bounded `Retry-After` value in both the header and response contract. UDP
overload and every invalid UDP input are silently dropped.
The same HTTP identity budget is computed whether or not a listing or attempt
exists. Rejection therefore does not disclose resource existence. Publisher
authentication also completes before any tenant/resource operation, while the
pre-authentication source budget prevents invalid credentials from bypassing
load shedding.
## Bounded state and recovery
`Rendezvous:AbuseProtection:MaxTrackedKeys` is a hard combined ceiling for rate
and active-concurrency keys. General HTTP and UDP traffic cannot consume the
configured `CriticalTrackedKeyReserve`; lease operations, health probes, and
allowlisted operator controls may
use that reserve but never exceed the hard ceiling. A request that would exceed
its applicable ceiling fails closed without adding state. Fixed-window rate keys
are cleared at the next window boundary; concurrency keys are removed as their
request leases finish. HTTP and UDP trackers have separate locks and cardinality
partitions, so a UDP flood cannot block HTTP admission on a shared lock or
consume HTTP key capacity. This gives
deterministic burst recovery and prevents an attacker from growing a permanent
high-cardinality address, credential, or resource table.
The complete default profile is checked into
`src/FinalFactory.Rendezvous.Server/appsettings.json`. Operators may lower or
tune limits for a measured deployment profile, but must preserve all dimensions
and leave the tracker ceiling above the maximum simultaneous key set. A rolling
deployment should use the same profile on every instance. These per-process
limits are a final service boundary; an edge proxy may add stricter distributed
limits but is not a substitute for them.
When an HTTP reverse proxy is used, every immediate proxy address must be
allowlisted in `Rendezvous:AbuseProtection:TrustedProxyAddresses` (or indexed
environment variables such as
`Rendezvous__AbuseProtection__TrustedProxyAddresses__0`). Only one forwarded
hop is accepted. With an empty allowlist, forwarded headers are ignored and the
direct TCP peer is the source. Never add a broad network range or accept
untrusted `X-Forwarded-For` input: that would let a caller choose its own rate
partition.
## Reflection, disclosure, and logging rules
- UDP sends nothing for malformed, oversized, unauthenticated, stale,
replayed, wrong-role, or rate-limited input.
- Introductions are emitted only after both role-scoped capabilities bind to
their observed gameplay-socket sources. HTTP never supplies a public
introduction target.
- Private candidates must be same-family private unicast addresses and are used
only for peers observed behind the same public address.
- Abuse keys, exceptions, and responses never include bearer credentials,
capabilities, tickets, raw endpoints, metadata values, or hostile markup.
- Endpoint and capability values are not used as metric labels or log fields.
## Verification
The deterministic test corpora use the recorded seeds `0x152026`, `0x154A50`,
and `0x1557A7E`. They exercise 10,000 arbitrary UDP envelopes through the
production decoder, 5,000 arbitrary HTTP/credential parser inputs, and 1,000
mutated state transitions, including the oversized and configured-capacity
boundaries.
Focused tests cover IPv4 and IPv6 prefix
partitioning, tenant/principal/resource concurrency, tracker exhaustion,
window recovery, wire-operation isolation, a steady-state allocation ceiling,
typed `429`/`413` responses, secret fingerprint redaction, and silent
authenticated UDP shedding. The existing state, contract, HTTP, client,
and mediator suites continue to cover cross-tenant access, replay, role swaps,
credential rotation, bounded metadata, endpoint validation, and one-shot
amplification behavior.
+1 -1
View File
@@ -11,7 +11,7 @@ backlog where the control is implemented and verified.
| Per-game credentials and signing keys | Provisioned principals and versioned keys are scoped to game/environment; secrets come from a provider and never a public binary. (#5) | Cross-tenant authorization tests, rotation/overlap/revocation tests, and secret scans. | | Per-game credentials and signing keys | Provisioned principals and versioned keys are scoped to game/environment; secrets come from a provider and never a public binary. (#5) | Cross-tenant authorization tests, rotation/overlap/revocation tests, and secret scans. |
| Short-lived, single-purpose tokens resistant to replay | Issuer fixes audience, tenant, attempt, role, issued/expiry times, nonce, and key ID; store atomically consumes nonce/ticket. (#4, #6, #10) | Golden vectors; expired, future, mutated, wrong-role, wrong-tenant, and concurrent replay tests. | | Short-lived, single-purpose tokens resistant to replay | Issuer fixes audience, tenant, attempt, role, issued/expiry times, nonce, and key ID; store atomically consumes nonce/ticket. (#4, #6, #10) | Golden vectors; expired, future, mutated, wrong-role, wrong-tenant, and concurrent replay tests. |
| Strict payload, metadata, and token size limits | ADR 0003 ceilings are checked before allocation/deserialization and again at domain construction. (#4, #15) | Boundary/property tests, malformed corpus, and allocation-aware fuzzing. | | Strict payload, metadata, and token size limits | ADR 0003 ceilings are checked before allocation/deserialization and again at domain construction. (#4, #15) | Boundary/property tests, malformed corpus, and allocation-aware fuzzing. |
| Registration, query, and introduction rate limits | Layered per-address, principal, tenant, and global token buckets with bounded queues and stable retry guidance. (#15) | Limit partition/isolation tests and overload/soak profiles. | | Registration, query, and introduction rate limits | Layered fixed-window budgets and concurrency leases cover global, operation, IPv4 `/24` or IPv6 `/56`, tenant, principal/capability, and listing/attempt dimensions with a bounded key table and stable retry guidance. (#15) | Deterministic partition, concurrency, tracker-exhaustion, recovery, typed-overload, and silent-UDP-shedding tests. |
| Lease expiry removes abandoned servers | Visibility and join eligibility atomically require a fresh lease and fresh authenticated presence. (#6, #7) | Fake-clock expiry, renew/expire race, restart, and stale-host join tests. | | Lease expiry removes abandoned servers | Visibility and join eligibility atomically require a fresh lease and fresh authenticated presence. (#6, #7) | Fake-clock expiry, renew/expire race, restart, and stale-host join tests. |
| Validate game, environment, room, and protocol boundaries | Every identifier is a validated type; store keys and authorization decisions include server-derived tenant scope; protocol is exact-match in v1. (#4-#10) | Contract, tenant-isolation, incompatible-version, and confused-deputy tests. | | Validate game, environment, room, and protocol boundaries | Every identifier is a validated type; store keys and authorization decisions include server-derived tenant scope; protocol is exact-match in v1. (#4-#10) | Contract, tenant-isolation, incompatible-version, and confused-deputy tests. |
| Structured audit events without secrets or reusable credentials | Allowlisted audit schema excludes metadata values, raw endpoints, tokens, and key material; event volume is bounded. (#16) | Captured-log/audit assertions and credential canary scans. | | Structured audit events without secrets or reusable credentials | Allowlisted audit schema excludes metadata values, raw endpoints, tokens, and key material; event volume is bounded. (#16) | Captured-log/audit assertions and credential canary scans. |
+7 -4
View File
@@ -63,8 +63,9 @@ only its public key ID/lifecycle metadata and does not require retired secret
material to remain available. material to remain available.
Key IDs are non-secret base64url identifiers. Secret references are resolved Key IDs are non-secret base64url identifiers. Secret references are resolved
through `ISecretProvider`; production supports `env:<VARIABLE>` references and through `ISecretProvider`; production supports base64 `env:<VARIABLE>` and raw
the interface is replaceable by a deployment-specific vault/KMS adapter. The `file:/absolute/path` references to bounded non-symlink files. The interface is
replaceable by a deployment-specific vault/KMS adapter. The
committed development profile uses an in-memory random key identified by a committed development profile uses an in-memory random key identified by a
`development:ephemeral/...` reference. It never writes key material to disk and `development:ephemeral/...` reference. It never writes key material to disk and
all credentials become invalid when the process exits. all credentials become invalid when the process exits.
@@ -74,8 +75,10 @@ all credentials become invalid when the process exits.
`Rendezvous:Provisioning` supplies issuer, audience, clock skew, signing-key `Rendezvous:Provisioning` supplies issuer, audience, clock skew, signing-key
descriptors, and game policies. A production key reference such as descriptors, and game policies. A production key reference such as
`env:RENDEZVOUS_SIGNING_KEY_2026_01` expects that environment variable to hold at `env:RENDEZVOUS_SIGNING_KEY_2026_01` expects that environment variable to hold at
least 32 random bytes encoded as base64. Missing, malformed, short, inactive, or least 32 random bytes encoded as base64. `file:/run/secrets/rendezvous-signing`
duplicate keys stop startup with a key-ID-only diagnostic. No game-wide secret expects the raw bytes in a read-only, absolute, non-symlink file. Missing,
malformed, short, inactive, or duplicate keys stop startup with a key-ID-only
diagnostic. No game-wide secret
belongs in `appsettings`, source control, examples, the Client package, URLs, belongs in `appsettings`, source control, examples, the Client package, URLs,
responses, logs, metrics, exceptions, or diagnostic dumps. responses, logs, metrics, exceptions, or diagnostic dumps.
+15
View File
@@ -0,0 +1,15 @@
<Project>
<PropertyGroup>
<RendezvousVersion>1.0.0</RendezvousVersion>
<RendezvousMajorVersion>1</RendezvousMajorVersion>
<RendezvousMinorVersion>0</RendezvousMinorVersion>
<RendezvousPatchVersion>0</RendezvousPatchVersion>
<MinimumClientVersion>1.0.0</MinimumClientVersion>
<MaximumClientMajorVersion>1</MaximumClientMajorVersion>
<HttpContractVersion>1</HttpContractVersion>
<UdpContractVersion>1</UdpContractVersion>
<ConnectionTicketFormatVersion>1</ConnectionTicketFormatVersion>
<LiteNetLibVersion>2.1.4</LiteNetLibVersion>
<LiteNetLibMajorVersion>2</LiteNetLibMajorVersion>
</PropertyGroup>
</Project>
+17
View File
@@ -0,0 +1,17 @@
{
"schemaVersion": 1,
"consumers": [
{
"name": "SpaceGame",
"repository": "https://git.finalfactory.de/Kyuubi/SpaceGame.git",
"revision": "77519b0cc418a27f8d408ae2d7b8812fbe087c04",
"project": "SpaceGame.csproj"
},
{
"name": "Unscouted",
"repository": "https://git.finalfactory.de/HeiKyu/Unscouted.git",
"revision": "7807dbee86eb8b98e702f1eb89c88adff728f635",
"project": "Net.Core/Net.Core.csproj"
}
]
}
+14
View File
@@ -0,0 +1,14 @@
# syntax=docker/dockerfile:1.7@sha256:a57df69d0ea827fb7266491f2813635de6f17269be881f696fbfdf2d83dda33e
FROM python:3.12.11-slim-bookworm@sha256:c00fc7b44d844b6da22861ec24af43968a5200eac4ec607b4725d585165d6b49 AS release-python
FROM ghcr.io/jqlang/jq:1.8.1@sha256:95de8f005ca027686a1ca3b0853e2bb219062438015862816159f3f25a4d4230 AS release-jq
FROM mcr.microsoft.com/dotnet/sdk:10.0.301-noble@sha256:ea8bde36c11b6e7eec2656d0e59101d4462f6bd630730f2c8201ed0572b295d5 AS release-builder
COPY --from=release-python /usr/local/ /usr/local/
COPY --from=release-jq /jq /usr/local/bin/jq
RUN dotnet --version \
&& python3 --version \
&& git --version \
&& tar --version \
&& gzip --version \
&& jq --version
WORKDIR /source
+27
View File
@@ -0,0 +1,27 @@
{
"schemaVersion": 1,
"packageRegistry": "https://git.finalfactory.de/api/packages/HeiKyu/nuget/index.json",
"containerRepository": "git.finalfactory.de/heikyu/rendezvous",
"allowedLicenseExpressions": [
"Apache-2.0",
"BSD-2-Clause",
"BSD-3-Clause",
"MIT"
],
"dependencyLicenseOverrides": {
"xunit.abstractions/2.0.3": {
"license": "Apache-2.0",
"reason": "Legacy package predates NuGet SPDX metadata; reviewed against the xUnit.net Apache-2.0 license."
}
},
"publishedPackages": [
"FinalFactory.Rendezvous.Client",
"FinalFactory.Rendezvous.Contracts"
],
"forbiddenArtifactNameFragments": [
"credential",
"password",
"private-key",
"signing-key"
]
}
+823
View File
@@ -0,0 +1,823 @@
#!/usr/bin/env python3
import argparse
import datetime as dt
import hashlib
import json
import os
import pathlib
import re
import sys
import zipfile
import xml.etree.ElementTree as ET
from xml.sax.saxutils import escape
PROJECT_PACKAGE_PREFIX = "finalfactory.rendezvous."
CORE_PROPERTIES_PATH = (
"package/services/metadata/core-properties/core-properties.psmdcp"
)
def fail(message: str) -> None:
raise SystemExit(message)
def load_json(path: pathlib.Path):
with path.open(encoding="utf-8") as stream:
return json.load(stream)
def dependency_inventory(root: pathlib.Path):
dependencies = {}
for lock_path in sorted(root.glob("**/packages.lock.json")):
if any(part in {"bin", "obj", "artifacts"} for part in lock_path.parts):
continue
lock = load_json(lock_path)
for framework in lock.get("dependencies", {}).values():
for package_id, details in framework.items():
resolved = details.get("resolved")
if not resolved or package_id.lower().startswith(PROJECT_PACKAGE_PREFIX):
continue
key = package_id.lower()
previous = dependencies.get(key)
if previous is not None and previous[1] != resolved:
fail(
f"Dependency {package_id} resolves to both {previous[1]} and {resolved}."
)
dependencies[key] = (package_id, resolved)
return [dependencies[key] for key in sorted(dependencies)]
def package_license(package_id: str, version: str, overrides):
package_root = pathlib.Path(
os.environ.get("NUGET_PACKAGES", pathlib.Path.home() / ".nuget" / "packages")
)
version_dir = package_root / package_id.lower() / version
nuspecs = list(version_dir.glob("*.nuspec"))
if len(nuspecs) != 1:
fail(f"Expected one restored nuspec for {package_id} {version} in {version_dir}.")
root = ET.parse(nuspecs[0]).getroot()
license_element = root.find(".//{*}license")
if license_element is None or license_element.get("type") != "expression":
override = overrides.get(f"{package_id.lower()}/{version}")
if override is None:
fail(f"{package_id} {version} does not declare an SPDX license expression.")
return override["license"]
expression = (license_element.text or "").strip()
if not expression:
fail(f"{package_id} {version} has an empty license expression.")
return expression
def command_policy(args) -> None:
root = pathlib.Path(args.root).resolve()
policy = load_json(root / "eng" / "release-policy.json")
allowed = set(policy["allowedLicenseExpressions"])
inventory = dependency_inventory(root)
observed = []
for package_id, version in inventory:
expression = package_license(
package_id, version, policy.get("dependencyLicenseOverrides", {})
)
if expression not in allowed:
fail(
f"Dependency {package_id} {version} uses unapproved license {expression}."
)
observed.append({"id": package_id, "version": version, "license": expression})
lite_net_lib = [item for item in observed if item["id"].lower() == "litenetlib"]
if lite_net_lib != [{"id": "LiteNetLib", "version": "2.1.4", "license": "MIT"}]:
fail(f"LiteNetLib must resolve exactly to the reviewed 2.1.4 release: {lite_net_lib}")
print(json.dumps({"dependencies": observed}, indent=2))
def command_audit(args) -> None:
document = load_json(pathlib.Path(args.input))
findings = []
def visit(value):
if isinstance(value, dict):
if value.get("vulnerabilities"):
findings.append(value)
for child in value.values():
visit(child)
elif isinstance(value, list):
for child in value:
visit(child)
visit(document)
if findings:
fail(f"Locked dependency graph contains known vulnerabilities: {findings}")
print("Locked dependency graph has no reported vulnerabilities.")
def release_dependency_graph(root: pathlib.Path, server_deps: pathlib.Path):
dependencies = {}
edges = set()
server_document = load_json(server_deps)
for package_key, details in server_document.get("libraries", {}).items():
if details.get("type") != "package":
continue
package_id, version = package_key.rsplit("/", 1)
dependencies[package_id.lower()] = (package_id, version)
server_target = next(iter(server_document.get("targets", {}).values()), {})
for source_key, details in server_target.items():
source_id = source_key.rsplit("/", 1)[0]
source = (
"SPDXRef-Server"
if source_id == "FinalFactory.Rendezvous.Server"
else source_id.lower()
)
for dependency_id in details.get("dependencies", {}):
target = {
"FinalFactory.Rendezvous.Contracts": "SPDXRef-Contracts",
"FinalFactory.Rendezvous.Client": "SPDXRef-Client",
}.get(dependency_id, dependency_id.lower())
edges.add((source, target))
lock_roots = (
("SPDXRef-Client", root / "src/FinalFactory.Rendezvous.Client/packages.lock.json"),
(
"SPDXRef-Contracts",
root / "src/FinalFactory.Rendezvous.Contracts/packages.lock.json",
),
)
for root_id, lock_path in lock_roots:
lock = load_json(lock_path)
for framework in lock.get("dependencies", {}).values():
for package_id, details in framework.items():
resolved = details.get("resolved")
if resolved and not package_id.lower().startswith(PROJECT_PACKAGE_PREFIX):
dependencies[package_id.lower()] = (package_id, resolved)
if details.get("type") == "Direct":
edges.add((root_id, package_id.lower()))
source = package_id.lower()
for dependency_id in details.get("dependencies", {}):
if dependency_id.lower().startswith(PROJECT_PACKAGE_PREFIX):
continue
edges.add((source, dependency_id.lower()))
edges.add(("SPDXRef-Client", "SPDXRef-Contracts"))
inventory = [dependencies[key] for key in sorted(dependencies)]
return inventory, edges
def command_sbom(args) -> None:
root = pathlib.Path(args.root).resolve()
policy = load_json(root / "eng" / "release-policy.json")
overrides = policy.get("dependencyLicenseOverrides", {})
packages = [
{
"SPDXID": "SPDXRef-Rendezvous",
"name": "FinalFactory.Rendezvous",
"versionInfo": args.version,
"downloadLocation": "NOASSERTION",
"filesAnalyzed": False,
"licenseConcluded": "NOASSERTION",
"licenseDeclared": "NOASSERTION",
"copyrightText": "NOASSERTION",
}
]
internal_packages = [
("SPDXRef-Server", "FinalFactory.Rendezvous.Server"),
("SPDXRef-Client", "FinalFactory.Rendezvous.Client"),
("SPDXRef-Contracts", "FinalFactory.Rendezvous.Contracts"),
]
for spdx_id, package_id in internal_packages:
packages.append(
{
"SPDXID": spdx_id,
"name": package_id,
"versionInfo": args.version,
"downloadLocation": "NOASSERTION",
"filesAnalyzed": False,
"licenseConcluded": "NOASSERTION",
"licenseDeclared": "NOASSERTION",
"copyrightText": "NOASSERTION",
}
)
inventory, dependency_edges = release_dependency_graph(
root, pathlib.Path(args.server_deps)
)
dependency_ids = {}
for index, (package_id, version) in enumerate(
inventory, start=1
):
expression = package_license(package_id, version, overrides)
spdx_id = f"SPDXRef-Package-{index}"
dependency_ids[package_id.lower()] = spdx_id
packages.append(
{
"SPDXID": spdx_id,
"name": package_id,
"versionInfo": version,
"downloadLocation": "NOASSERTION",
"filesAnalyzed": False,
"licenseConcluded": expression,
"licenseDeclared": expression,
"copyrightText": "NOASSERTION",
"externalRefs": [
{
"referenceCategory": "PACKAGE-MANAGER",
"referenceType": "purl",
"referenceLocator": f"pkg:nuget/{package_id}@{version}",
}
],
}
)
created = dt.datetime.fromtimestamp(
int(args.source_date_epoch), dt.timezone.utc
).strftime("%Y-%m-%dT%H:%M:%SZ")
document = {
"spdxVersion": "SPDX-2.3",
"dataLicense": "CC0-1.0",
"SPDXID": "SPDXRef-DOCUMENT",
"name": f"FinalFactory.Rendezvous-{args.version}",
"documentNamespace": (
"https://git.finalfactory.de/HeiKyu/Rendezvous/sbom/"
f"{args.version}/{args.commit}"
),
"creationInfo": {
"created": created,
"creators": ["Tool: eng/release_artifacts.py"],
},
"documentDescribes": ["SPDXRef-Rendezvous"],
"packages": packages,
"relationships": [
{
"spdxElementId": "SPDXRef-Rendezvous",
"relationshipType": "CONTAINS",
"relatedSpdxElement": spdx_id,
}
for spdx_id, _ in internal_packages
]
+ [
{
"spdxElementId": dependency_ids.get(source, source),
"relationshipType": "DEPENDS_ON",
"relatedSpdxElement": dependency_ids.get(target, target),
}
for source, target in sorted(dependency_edges)
if source in dependency_ids or source.startswith("SPDXRef-")
if target in dependency_ids or target.startswith("SPDXRef-")
],
}
output = pathlib.Path(args.output)
output.parent.mkdir(parents=True, exist_ok=True)
output.write_text(json.dumps(document, indent=2) + "\n", encoding="utf-8")
def nuspec_metadata(archive: pathlib.Path):
with zipfile.ZipFile(archive) as package:
names = package.namelist()
nuspecs = [name for name in names if name.endswith(".nuspec")]
if len(nuspecs) != 1:
fail(f"{archive.name} must contain exactly one nuspec.")
root = ET.fromstring(package.read(nuspecs[0]))
metadata = root.find(".//{*}metadata")
if metadata is None:
fail(f"{archive.name} has no package metadata.")
values = {
child.tag.rsplit("}", 1)[-1]: (child.text or "").strip()
for child in metadata
if len(child) == 0
}
dependencies = {
item.get("id"): item.get("version")
for item in metadata.findall(".//{*}dependency")
}
repository = metadata.find("./{*}repository")
repository_attributes = {} if repository is None else dict(repository.attrib)
return values, dependencies, repository_attributes
def verify_checksum_file(release_dir: pathlib.Path, excluded=()) -> None:
checksum_path = release_dir / "checksums.sha256"
lines = checksum_path.read_text(encoding="utf-8").splitlines()
if not lines:
fail("checksums.sha256 is empty.")
referenced = set()
for line in lines:
digest, marker, relative = line.partition(" ")
if marker != " " or not re.fullmatch(r"[0-9a-f]{64}", digest):
fail(f"Malformed checksum line: {line}")
target = release_dir / relative
if not target.is_file():
fail(f"Checksum references missing artifact: {relative}")
actual = hashlib.sha256(target.read_bytes()).hexdigest()
if actual != digest:
fail(f"Checksum mismatch for {relative}.")
referenced.add(relative)
expected = {
path.name
for path in release_dir.iterdir()
if path.is_file()
and path.name != "checksums.sha256"
and path.name not in excluded
}
if referenced != expected:
fail(
"Checksum manifest coverage differs. "
f"Missing={expected - referenced}; extra={referenced - expected}"
)
def command_verify(args) -> None:
release_dir = pathlib.Path(args.release_dir).resolve()
version = args.version
expected = {
f"FinalFactory.Rendezvous.Client.{version}.nupkg",
f"FinalFactory.Rendezvous.Client.{version}.snupkg",
f"FinalFactory.Rendezvous.Contracts.{version}.nupkg",
f"FinalFactory.Rendezvous.Contracts.{version}.snupkg",
f"FinalFactory.Rendezvous.Server.{version}.linux-x64.tar.gz",
f"FinalFactory.Rendezvous.{version}.spdx.json",
"CHANGELOG.md",
"checksums.sha256",
"release-provenance.json",
}
checksum_exclusions = set()
if args.phase in {"publish-ready", "signing-ready", "published"}:
expected.add(f"FinalFactory.Rendezvous.Container.{version}.spdx.json")
if args.phase in {"signing-ready", "published"}:
expected.update({"container-digest.txt", "cosign.pub"})
if args.phase == "published":
expected.add("checksums.sha256.bundle")
checksum_exclusions.add("checksums.sha256.bundle")
actual = {path.name for path in release_dir.iterdir() if path.is_file()}
if actual != expected:
fail(f"Release artifact set differs. Missing={expected - actual}; extra={actual - expected}")
if args.phase in {"publish-ready", "signing-ready", "published"}:
container_sbom = load_json(
release_dir / f"FinalFactory.Rendezvous.Container.{version}.spdx.json"
)
if container_sbom.get("spdxVersion") != "SPDX-2.3":
fail("Container inventory must be an SPDX 2.3 document.")
if not container_sbom.get("packages"):
fail("Container SPDX inventory contains no packages.")
policy = load_json(pathlib.Path(args.root) / "eng" / "release-policy.json")
forbidden = tuple(fragment.lower() for fragment in policy["forbiddenArtifactNameFragments"])
for artifact in actual:
if artifact != "release-provenance.json" and any(fragment in artifact.lower() for fragment in forbidden):
fail(f"Forbidden secret-like artifact name: {artifact}")
release_sbom = load_json(
release_dir / f"FinalFactory.Rendezvous.{version}.spdx.json"
)
package_ids = {
package.get("name"): package.get("SPDXID")
for package in release_sbom.get("packages", [])
}
relationships = {
(
relationship.get("spdxElementId"),
relationship.get("relationshipType"),
relationship.get("relatedSpdxElement"),
)
for relationship in release_sbom.get("relationships", [])
}
expected_component_edges = {
("SPDXRef-Server", "SPDXRef-Contracts"),
("SPDXRef-Server", package_ids.get("LiteNetLib")),
("SPDXRef-Server", package_ids.get("Microsoft.AspNetCore.OpenApi")),
("SPDXRef-Client", "SPDXRef-Contracts"),
("SPDXRef-Client", package_ids.get("LiteNetLib")),
("SPDXRef-Contracts", package_ids.get("System.Text.Json")),
}
for source, target in expected_component_edges:
if not target or (source, "DEPENDS_ON", target) not in relationships:
fail(f"Release SPDX inventory is missing component edge {source} -> {target}.")
bcl_id = package_ids.get("Microsoft.Bcl.AsyncInterfaces")
if ("SPDXRef-Server", "DEPENDS_ON", bcl_id) in relationships:
fail("Release SPDX inventory incorrectly flattens transitive dependencies onto Server.")
for package_id in policy["publishedPackages"]:
package = release_dir / f"{package_id}.{version}.nupkg"
metadata, dependencies, repository = nuspec_metadata(package)
if metadata.get("id") != package_id or metadata.get("version") != version:
fail(f"{package.name} identity/version metadata is incorrect.")
if metadata.get("projectUrl") != "https://git.finalfactory.de/HeiKyu/Rendezvous":
fail(f"{package.name} has an incorrect project URL.")
if repository.get("url") != "https://git.finalfactory.de/HeiKyu/Rendezvous":
fail(f"{package.name} has an incorrect repository URL.")
if repository.get("commit") != provenance_commit(release_dir):
fail(f"{package.name} does not identify the release commit.")
symbol_package = release_dir / f"{package_id}.{version}.snupkg"
with zipfile.ZipFile(symbol_package) as symbols:
if not any(name.endswith(".pdb") for name in symbols.namelist()):
fail(f"{symbol_package.name} contains no portable PDB.")
with zipfile.ZipFile(package) as archive:
names = set(archive.namelist())
required_entries = {
"README.md",
"CHANGELOG.md",
f"lib/netstandard2.1/{package_id}.dll",
}
if not required_entries <= names:
fail(
f"{package.name} is missing required package content: "
f"{required_entries - names}"
)
forbidden_entries = [
name
for name in names
if any(
fragment in name.lower()
for fragment in (
"appsettings",
"launchsettings",
".env",
"credential",
"signing-key",
"private-key",
)
)
]
if forbidden_entries:
fail(
f"{package.name} contains deployable configuration or secrets: "
f"{forbidden_entries}"
)
if package_id.endswith(".Client"):
if dependencies.get("LiteNetLib") != "[2.1.4]":
fail(f"Client package must pin LiteNetLib exactly to 2.1.4: {dependencies}")
contracts_range = dependencies.get("FinalFactory.Rendezvous.Contracts", "")
if contracts_range != f"[{version}]":
fail(f"Client package does not depend on the matching Contracts version.")
provenance = load_json(release_dir / "release-provenance.json")
if provenance.get("version") != version:
fail("Release provenance does not identify the requested version.")
if provenance.get("treeState") != "clean" and not args.allow_dirty:
fail("Release provenance must identify a clean tree.")
container = provenance.get("containerImage", "")
if container.endswith(":latest") or ":latest@" in container or f":{version}" not in container:
fail(f"Container reference is mutable or not versioned: {container}")
if provenance.get("containerPlatform") != "linux/amd64":
fail("Release provenance must pin the linux/amd64 container platform.")
for field in ("containerBaseDigests", "releaseBuilderBaseDigests"):
images = provenance.get(field, [])
if not images or any("@sha256:" not in image or image.endswith(":latest") for image in images):
fail(f"Release provenance contains an unpinned build image in {field}: {images}")
build_tools = provenance.get("buildTools", [])
required_tool_prefixes = ("dotnet", "python=", "tar=", "gzip=", "jq=")
observed_tools = [f"dotnet={provenance.get('dotnetSdk', '')}", *build_tools]
for prefix in required_tool_prefixes:
if not any(tool.startswith(prefix) for tool in observed_tools):
fail(f"Release provenance is missing an artifact tool version: {prefix}")
if args.phase in {"publish-ready", "signing-ready", "published"}:
if not re.fullmatch(
r"sha256:[0-9a-f]{64}", provenance.get("containerImageId", "")
):
fail("Release provenance is missing the verified local container image ID.")
expected_namespace = (
"https://git.finalfactory.de/HeiKyu/Rendezvous/container-sbom/"
f"{version}/{provenance_commit(release_dir)}"
)
if container_sbom.get("documentNamespace") != expected_namespace:
fail("Container SPDX inventory does not identify the release commit.")
expected_created = dt.datetime.fromtimestamp(
int(provenance.get("sourceDateEpoch", 0)), dt.timezone.utc
).strftime("%Y-%m-%dT%H:%M:%SZ")
if container_sbom.get("creationInfo", {}).get("created") != expected_created:
fail("Container SPDX timestamp is not normalized to the source epoch.")
if args.phase in {"signing-ready", "published"}:
digest = (release_dir / "container-digest.txt").read_text(
encoding="utf-8"
).strip()
if not re.fullmatch(
r"git\.finalfactory\.de/heikyu/rendezvous@sha256:[0-9a-f]{64}", digest
):
fail(f"Published container digest is invalid: {digest}")
if provenance.get("containerDigest") != digest:
fail("Published provenance and container digest file differ.")
for prefix in ("docker-buildx=", "buildkit="):
if not any(tool.startswith(prefix) for tool in build_tools):
fail(f"Published provenance is missing container tool version: {prefix}")
verify_checksum_file(release_dir, checksum_exclusions)
print(f"Verified {args.phase} release artifact set for {version}.")
def provenance_commit(release_dir: pathlib.Path) -> str:
provenance = load_json(release_dir / "release-provenance.json")
commit = provenance.get("commit", "")
if not re.fullmatch(r"[0-9a-f]{40}", commit):
fail("Release provenance must contain a full Git commit SHA.")
return commit
def command_consumer(args) -> None:
assets = load_json(pathlib.Path(args.assets))
libraries = assets.get("libraries", {})
required = {
f"FinalFactory.Rendezvous.Client/{args.version}",
f"FinalFactory.Rendezvous.Contracts/{args.version}",
"LiteNetLib/2.1.4",
}
missing = required - set(libraries)
if missing:
fail(f"Consumer restore is missing exact release dependencies: {missing}")
forbidden = [name for name in libraries if name.lower().startswith("litenetlib/1.")]
if forbidden:
fail(f"Consumer resolved forbidden LiteNetLib 1.x assets: {forbidden}")
def command_consumer_config(args) -> None:
local_source = escape(str(pathlib.Path(args.local_source).resolve()))
configuration = f'''<?xml version="1.0" encoding="utf-8"?>
<configuration>
<packageSources>
<clear />
<add key="rendezvous-candidate" value="{local_source}" />
<add key="nuget.org" value="https://api.nuget.org/v3/index.json" protocolVersion="3" />
</packageSources>
<packageSourceMapping>
<packageSource key="rendezvous-candidate">
<package pattern="FinalFactory.Rendezvous.*" />
</packageSource>
<packageSource key="nuget.org">
<package pattern="*" />
</packageSource>
</packageSourceMapping>
</configuration>
'''
pathlib.Path(args.output).write_text(configuration, encoding="utf-8")
def command_source_link(args) -> None:
document = load_json(pathlib.Path(args.file))
mappings = document.get("documents", {})
expected = (
"https://git.finalfactory.de/HeiKyu/Rendezvous/raw/commit/"
f"{args.commit}/"
)
if not mappings or any(not value.startswith(expected) for value in mappings.values()):
fail(f"SourceLink mappings do not identify commit {args.commit}: {mappings}")
def command_normalize_package(args) -> None:
package_path = pathlib.Path(args.package).resolve()
if package_path.suffix not in {".nupkg", ".snupkg"}:
fail(f"Unsupported NuGet archive extension: {package_path.name}")
timestamp = dt.datetime.fromtimestamp(
int(args.source_date_epoch), dt.timezone.utc
)
zip_timestamp = (
max(timestamp.year, 1980),
timestamp.month,
timestamp.day,
timestamp.hour,
timestamp.minute,
timestamp.second - (timestamp.second % 2),
)
with zipfile.ZipFile(package_path) as source:
entries = {name: source.read(name) for name in source.namelist()}
if ".signature.p7s" in entries:
fail(f"Refusing to rewrite signed package: {package_path.name}")
core_paths = [
name
for name in entries
if name.startswith("package/services/metadata/core-properties/")
and name.endswith(".psmdcp")
]
if len(core_paths) != 1:
fail(f"Expected one NuGet core-properties part in {package_path.name}.")
entries[CORE_PROPERTIES_PATH] = entries.pop(core_paths[0])
nuspec_paths = [name for name in entries if name.endswith(".nuspec")]
if len(nuspec_paths) != 1:
fail(f"Expected one nuspec in {package_path.name}.")
nuspec = ET.fromstring(entries[nuspec_paths[0]])
nuspec_namespace = nuspec.tag.partition("}")[0].removeprefix("{")
if nuspec_namespace:
ET.register_namespace("", nuspec_namespace)
package_id = nuspec.findtext(".//{*}id")
if package_id == "FinalFactory.Rendezvous.Client":
contracts = nuspec.find(
".//{*}dependency[@id='FinalFactory.Rendezvous.Contracts']"
)
if contracts is None:
fail("Client package has no Contracts dependency to pin.")
contracts.set("version", f"[{args.version}]")
entries[nuspec_paths[0]] = ET.tostring(
nuspec, encoding="utf-8", xml_declaration=True
)
relationships_namespace = (
"http://schemas.openxmlformats.org/package/2006/relationships"
)
ET.register_namespace("", relationships_namespace)
relationships = ET.fromstring(entries["_rels/.rels"])
for relationship in relationships:
relationship_type = relationship.get("Type", "")
if relationship_type.endswith("/manifest"):
relationship.set("Id", "RManifest")
elif relationship_type.endswith("/metadata/core-properties"):
relationship.set("Id", "RCoreProperties")
relationship.set("Target", f"/{CORE_PROPERTIES_PATH}")
entries["_rels/.rels"] = ET.tostring(
relationships, encoding="utf-8", xml_declaration=True
)
temporary = package_path.with_suffix(package_path.suffix + ".normalized")
with zipfile.ZipFile(temporary, "w", compression=zipfile.ZIP_STORED) as target:
for name in sorted(entries):
info = zipfile.ZipInfo(name, date_time=zip_timestamp)
info.compress_type = zipfile.ZIP_STORED
info.create_system = 3
info.external_attr = 0o100644 << 16
target.writestr(info, entries[name])
temporary.replace(package_path)
def command_provenance(args) -> None:
versions = ET.parse(pathlib.Path(args.root) / "eng" / "Versions.props")
def version_property(name: str) -> str:
element = versions.find(f".//{name}")
if element is None or not element.text:
fail(f"Missing central release property: {name}")
return element.text.strip()
provenance = {
"schemaVersion": 1,
"version": args.version,
"commit": args.commit,
"treeState": args.tree_state,
"buildConfiguration": "Release",
"sourceDateEpoch": int(args.source_date_epoch),
"dotnetSdk": args.dotnet_sdk,
"buildTools": sorted(args.build_tool),
"containerImage": f"git.finalfactory.de/heikyu/rendezvous:{args.version}",
"containerPlatform": "linux/amd64",
"containerBaseDigests": args.base_digest,
"releaseBuilderBaseDigests": args.builder_base,
"packages": [
f"FinalFactory.Rendezvous.Client/{args.version}",
f"FinalFactory.Rendezvous.Contracts/{args.version}",
],
"compatibility": {
"minimumClientVersion": version_property("MinimumClientVersion"),
"maximumClientMajorVersion": int(version_property("MaximumClientMajorVersion")),
"httpContractVersions": [int(version_property("HttpContractVersion"))],
"udpContractVersions": [int(version_property("UdpContractVersion"))],
"connectionTicketFormatVersions": [
int(version_property("ConnectionTicketFormatVersion"))
],
"liteNetLib": version_property("LiteNetLibVersion"),
"gameplayProtocol": "exact-per-tenant",
},
}
pathlib.Path(args.output).write_text(
json.dumps(provenance, indent=2) + "\n", encoding="utf-8"
)
def command_record_container_build(args) -> None:
path = pathlib.Path(args.provenance)
provenance = load_json(path)
tools = set(provenance.get("buildTools", []))
tools.add(f"docker-buildx={args.buildx_version}")
tools.add(f"buildkit={args.buildkit_version}")
provenance["buildTools"] = sorted(tools)
provenance["containerPlatform"] = "linux/amd64"
if not re.fullmatch(r"sha256:[0-9a-f]{64}", args.image_id):
fail(f"Container image ID is invalid: {args.image_id}")
provenance["containerImageId"] = args.image_id
path.write_text(json.dumps(provenance, indent=2) + "\n", encoding="utf-8")
def command_record_container_digest(args) -> None:
if not re.fullmatch(
r"git\.finalfactory\.de/heikyu/rendezvous@sha256:[0-9a-f]{64}",
args.digest,
):
fail(f"Published container digest is invalid: {args.digest}")
release_dir = pathlib.Path(args.release_dir)
provenance_path = release_dir / "release-provenance.json"
provenance = load_json(provenance_path)
provenance["containerDigest"] = args.digest
provenance_path.write_text(
json.dumps(provenance, indent=2) + "\n", encoding="utf-8"
)
(release_dir / "container-digest.txt").write_text(
args.digest + "\n", encoding="utf-8"
)
def command_normalize_container_sbom(args) -> None:
path = pathlib.Path(args.file)
document = load_json(path)
created = dt.datetime.fromtimestamp(
int(args.source_date_epoch), dt.timezone.utc
).strftime("%Y-%m-%dT%H:%M:%SZ")
document["name"] = f"FinalFactory.Rendezvous.Container-{args.version}"
document["documentNamespace"] = (
"https://git.finalfactory.de/HeiKyu/Rendezvous/container-sbom/"
f"{args.version}/{args.commit}"
)
creation = document.setdefault("creationInfo", {})
creation["created"] = created
creation["creators"] = ["Tool: Trivy-0.69.3"]
if isinstance(document.get("packages"), list):
document["packages"] = sorted(
document["packages"],
key=lambda item: (
item.get("SPDXID", ""),
item.get("name", ""),
item.get("versionInfo", ""),
),
)
if isinstance(document.get("relationships"), list):
document["relationships"] = sorted(
document["relationships"],
key=lambda item: (
item.get("spdxElementId", ""),
item.get("relationshipType", ""),
item.get("relatedSpdxElement", ""),
),
)
path.write_text(
json.dumps(document, indent=2, sort_keys=True) + "\n", encoding="utf-8"
)
def main() -> None:
parser = argparse.ArgumentParser()
subparsers = parser.add_subparsers(dest="command", required=True)
policy = subparsers.add_parser("policy")
policy.add_argument("--root", required=True)
policy.set_defaults(handler=command_policy)
audit = subparsers.add_parser("audit")
audit.add_argument("--input", required=True)
audit.set_defaults(handler=command_audit)
sbom = subparsers.add_parser("sbom")
sbom.add_argument("--root", required=True)
sbom.add_argument("--version", required=True)
sbom.add_argument("--commit", required=True)
sbom.add_argument("--source-date-epoch", required=True)
sbom.add_argument("--server-deps", required=True)
sbom.add_argument("--output", required=True)
sbom.set_defaults(handler=command_sbom)
verify = subparsers.add_parser("verify")
verify.add_argument("--root", required=True)
verify.add_argument("--release-dir", required=True)
verify.add_argument("--version", required=True)
verify.add_argument("--allow-dirty", action="store_true")
verify.add_argument(
"--phase",
choices=("build", "publish-ready", "signing-ready", "published"),
default="build",
)
verify.set_defaults(handler=command_verify)
consumer = subparsers.add_parser("consumer")
consumer.add_argument("--assets", required=True)
consumer.add_argument("--version", required=True)
consumer.set_defaults(handler=command_consumer)
consumer_config = subparsers.add_parser("consumer-config")
consumer_config.add_argument("--local-source", required=True)
consumer_config.add_argument("--output", required=True)
consumer_config.set_defaults(handler=command_consumer_config)
source_link = subparsers.add_parser("source-link")
source_link.add_argument("--file", required=True)
source_link.add_argument("--commit", required=True)
source_link.set_defaults(handler=command_source_link)
normalize = subparsers.add_parser("normalize-package")
normalize.add_argument("--package", required=True)
normalize.add_argument("--source-date-epoch", required=True)
normalize.add_argument("--version", required=True)
normalize.set_defaults(handler=command_normalize_package)
provenance = subparsers.add_parser("provenance")
provenance.add_argument("--root", required=True)
provenance.add_argument("--version", required=True)
provenance.add_argument("--commit", required=True)
provenance.add_argument("--tree-state", required=True)
provenance.add_argument("--source-date-epoch", required=True)
provenance.add_argument("--dotnet-sdk", required=True)
provenance.add_argument("--build-tool", action="append", default=[])
provenance.add_argument("--base-digest", action="append", default=[])
provenance.add_argument("--builder-base", action="append", default=[])
provenance.add_argument("--output", required=True)
provenance.set_defaults(handler=command_provenance)
record_container = subparsers.add_parser("record-container-build")
record_container.add_argument("--provenance", required=True)
record_container.add_argument("--buildx-version", required=True)
record_container.add_argument("--buildkit-version", required=True)
record_container.add_argument("--image-id", required=True)
record_container.set_defaults(handler=command_record_container_build)
record_digest = subparsers.add_parser("record-container-digest")
record_digest.add_argument("--release-dir", required=True)
record_digest.add_argument("--digest", required=True)
record_digest.set_defaults(handler=command_record_container_digest)
normalize_container_sbom = subparsers.add_parser("normalize-container-sbom")
normalize_container_sbom.add_argument("--file", required=True)
normalize_container_sbom.add_argument("--version", required=True)
normalize_container_sbom.add_argument("--commit", required=True)
normalize_container_sbom.add_argument("--source-date-epoch", required=True)
normalize_container_sbom.set_defaults(handler=command_normalize_container_sbom)
args = parser.parse_args()
args.handler(args)
if __name__ == "__main__":
main()
+253
View File
@@ -0,0 +1,253 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:-}"
output="${2:-}"
property() {
sed -n "s:.*<$1>\(.*\)</$1>.*:\1:p" "$root/eng/Versions.props"
}
if [[ -z "$version" ]]; then
version="$(property RendezvousVersion)"
fi
semver='^(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)(-([0-9A-Za-z-]+\.)*[0-9A-Za-z-]+)?(\+([0-9A-Za-z-]+\.)*[0-9A-Za-z-]+)?$'
if [[ ! "$version" =~ $semver ]]; then
echo "Release version is not valid SemVer: $version" >&2
exit 1
fi
prerelease="${version%%+*}"
if [[ "$prerelease" == *-* ]]; then
prerelease="${prerelease#*-}"
IFS='.' read -r -a prerelease_identifiers <<<"$prerelease"
for identifier in "${prerelease_identifiers[@]}"; do
if [[ "$identifier" =~ ^[0-9]+$ && ! "$identifier" =~ ^(0|[1-9][0-9]*)$ ]]; then
echo "Numeric prerelease identifiers must not contain leading zeroes: $version" >&2
exit 1
fi
done
fi
if [[ "$version" != "$(property RendezvousVersion)" ]]; then
echo "Requested version $version differs from eng/Versions.props." >&2
exit 1
fi
for command in dotnet git python3 tar gzip sha256sum cmp jq; do
command -v "$command" >/dev/null || {
echo "Required release command is unavailable: $command" >&2
exit 1
}
done
commit="$(git -C "$root" rev-parse HEAD)"
source_date_epoch="$(git -C "$root" show -s --format=%ct "$commit")"
tree_state=clean
if [[ -n "$(git -C "$root" status --porcelain --untracked-files=normal)" ]]; then
tree_state=dirty
fi
allow_dirty=()
if [[ "$tree_state" != clean ]]; then
if [[ "${RENDEZVOUS_RELEASE_ALLOW_DIRTY:-0}" != 1 ]]; then
echo "A formal release must be built from a clean Git tree." >&2
exit 1
fi
allow_dirty=(--allow-dirty)
fi
if [[ -z "$output" ]]; then
output="$root/artifacts/release/$version"
fi
if [[ -e "$output" ]]; then
echo "Release output already exists; refusing to overwrite: $output" >&2
exit 1
fi
mkdir -p "$(dirname "$output")"
output="$(cd "$(dirname "$output")" && pwd)/$(basename "$output")"
work="$(mktemp -d "${TMPDIR:-/tmp}/rendezvous-release.XXXXXX")"
cleanup() {
rm -rf "$work"
}
trap cleanup EXIT
mkdir -p "$work/pack-1" "$work/pack-2" "$work/publish-1" "$work/publish-2" "$output"
create_server_archive() {
local publish_directory="$1"
local destination="$2"
tar --sort=name \
--mtime="@$source_date_epoch" \
--owner=0 --group=0 --numeric-owner \
-C "$publish_directory" -cf - . \
| gzip -n >"$destination"
}
common=(
-p:ContinuousIntegrationBuild=true
-p:PackageVersion="$version"
-p:RepositoryCommit="$commit"
-p:SourceRevisionId="$commit"
)
cd "$root"
dotnet restore Rendezvous.slnx --locked-mode
python3 eng/release_artifacts.py policy --root "$root" >"$work/dependency-policy.json"
dotnet package list \
--project Rendezvous.slnx \
--vulnerable \
--include-transitive \
--no-restore \
--format json >"$work/nuget-vulnerabilities.json"
python3 eng/release_artifacts.py audit --input "$work/nuget-vulnerabilities.json"
dotnet format Rendezvous.slnx --verify-no-changes --no-restore
api_before="$(sha256sum docs/api/*.json)"
dotnet build Rendezvous.slnx --configuration Release --no-restore "${common[@]}"
cp src/FinalFactory.Rendezvous.Server/bin/Release/net10.0/FinalFactory.Rendezvous.Server.dll \
"$work/FinalFactory.Rendezvous.Server.first.dll"
cp src/FinalFactory.Rendezvous.Server/bin/Release/net10.0/FinalFactory.Rendezvous.Server.pdb \
"$work/FinalFactory.Rendezvous.Server.first.pdb"
dotnet publish src/FinalFactory.Rendezvous.Server/FinalFactory.Rendezvous.Server.csproj \
--configuration Release \
--no-build \
--no-restore \
--output "$work/publish-1" \
-p:UseAppHost=false \
-p:OpenApiGenerateDocuments=false \
"${common[@]}"
create_server_archive "$work/publish-1" "$work/FinalFactory.Rendezvous.Server.first.tar.gz"
if [[ "$tree_state" == clean ]]; then
git diff --exit-code -- docs/api
elif [[ "$api_before" != "$(sha256sum docs/api/*.json)" ]]; then
echo "Generated OpenAPI changed during the release build." >&2
exit 1
fi
dotnet test Rendezvous.slnx --configuration Release --no-build
for project in Client Contracts; do
dotnet pack "src/FinalFactory.Rendezvous.$project/FinalFactory.Rendezvous.$project.csproj" \
--configuration Release --no-build --output "$work/pack-1" "${common[@]}"
done
for package in "$work/pack-1"/*; do
python3 eng/release_artifacts.py normalize-package \
--package "$package" \
--source-date-epoch "$source_date_epoch" \
--version "$version"
done
# Rebuild from the locked graph and prove package byte reproducibility.
dotnet clean Rendezvous.slnx --configuration Release >/dev/null
dotnet restore Rendezvous.slnx --locked-mode
dotnet build Rendezvous.slnx --configuration Release --no-restore "${common[@]}"
for project in Client Contracts; do
python3 eng/release_artifacts.py source-link \
--file "src/FinalFactory.Rendezvous.$project/obj/Release/netstandard2.1/FinalFactory.Rendezvous.$project.sourcelink.json" \
--commit "$commit"
done
cmp --silent \
"$work/FinalFactory.Rendezvous.Server.first.dll" \
src/FinalFactory.Rendezvous.Server/bin/Release/net10.0/FinalFactory.Rendezvous.Server.dll || {
echo "Server assembly is not byte reproducible." >&2
exit 1
}
cmp --silent \
"$work/FinalFactory.Rendezvous.Server.first.pdb" \
src/FinalFactory.Rendezvous.Server/bin/Release/net10.0/FinalFactory.Rendezvous.Server.pdb || {
echo "Server portable PDB is not byte reproducible." >&2
exit 1
}
for project in Client Contracts; do
dotnet pack "src/FinalFactory.Rendezvous.$project/FinalFactory.Rendezvous.$project.csproj" \
--configuration Release --no-build --output "$work/pack-2" "${common[@]}"
done
for package in "$work/pack-2"/*; do
python3 eng/release_artifacts.py normalize-package \
--package "$package" \
--source-date-epoch "$source_date_epoch" \
--version "$version"
done
for package in "$work/pack-1"/*; do
cmp --silent "$package" "$work/pack-2/$(basename "$package")" || {
echo "Package is not byte reproducible: $(basename "$package")" >&2
exit 1
}
done
cp "$work/pack-1"/* "$output/"
python3 eng/release_artifacts.py consumer-config \
--local-source "$output" \
--output "$work/consumer.NuGet.config"
for consumer in spacegame unscouted; do
project="$root/tests/consumers/$consumer/$(find "$root/tests/consumers/$consumer" -maxdepth 1 -name '*.csproj' -printf '%f\n')"
packages="$work/consumer-packages-$consumer"
dotnet restore "$project" \
-p:RendezvousPackageVersion="$version" \
-p:RestoreLockedMode=false \
--packages "$packages" \
--configfile "$work/consumer.NuGet.config" \
--force-evaluate
dotnet build "$project" \
--configuration Release --no-restore \
-p:RendezvousPackageVersion="$version"
assets="$(dirname "$project")/obj/project.assets.json"
python3 eng/release_artifacts.py consumer --assets "$assets" --version "$version"
done
dotnet publish src/FinalFactory.Rendezvous.Server/FinalFactory.Rendezvous.Server.csproj \
--configuration Release \
--no-build \
--no-restore \
--output "$work/publish-2" \
-p:UseAppHost=false \
-p:OpenApiGenerateDocuments=false \
"${common[@]}"
server_archive="$output/FinalFactory.Rendezvous.Server.$version.linux-x64.tar.gz"
create_server_archive "$work/publish-2" "$server_archive"
cmp --silent "$work/FinalFactory.Rendezvous.Server.first.tar.gz" "$server_archive" || {
echo "Server archive is not byte reproducible." >&2
exit 1
}
cp CHANGELOG.md "$output/CHANGELOG.md"
python3 eng/release_artifacts.py sbom \
--root "$root" \
--version "$version" \
--commit "$commit" \
--source-date-epoch "$source_date_epoch" \
--server-deps "$work/publish-2/FinalFactory.Rendezvous.Server.deps.json" \
--output "$output/FinalFactory.Rendezvous.$version.spdx.json"
provenance=(
provenance
--root "$root"
--version "$version"
--commit "$commit"
--tree-state "$tree_state"
--source-date-epoch "$source_date_epoch"
--dotnet-sdk "$(dotnet --version)"
--build-tool "python=$(python3 --version 2>&1)"
--build-tool "tar=$(tar --version | sed -n '1p')"
--build-tool "gzip=$(gzip --version | sed -n '1p')"
--build-tool "jq=$(jq --version)"
--output "$output/release-provenance.json"
)
while IFS= read -r base; do
provenance+=(--base-digest "$base")
done < <(sed -n 's/^FROM \([^ ]*\).*/\1/p' Dockerfile)
while IFS= read -r base; do
provenance+=(--builder-base "$base")
done < <(sed -n 's/^FROM \([^ ]*\).*/\1/p' eng/release-builder.Dockerfile)
python3 eng/release_artifacts.py "${provenance[@]}"
(
cd "$output"
find . -maxdepth 1 -type f ! -name checksums.sha256 -printf '%f\n' \
| LC_ALL=C sort \
| xargs sha256sum >checksums.sha256
)
python3 eng/release_artifacts.py verify \
--root "$root" \
--release-dir "$output" \
--version "$version" \
"${allow_dirty[@]}"
echo "Release artifacts verified at $output"
+56
View File
@@ -0,0 +1,56 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
base="${1:-origin/main}"
if ! git -C "$root" cat-file -e "$base^{commit}" 2>/dev/null; then
echo "Compatibility base does not exist; this is valid only for the initial version baseline: $base"
exit 0
fi
if [[ "$(git -C "$root" rev-parse "$base")" == "$(git -C "$root" rev-parse HEAD)" ]]; then
base="HEAD^"
fi
if ! git -C "$root" cat-file -e "$base:eng/Versions.props" 2>/dev/null; then
echo "Base has no release version manifest; accepting the initial compatibility baseline."
exit 0
fi
current_property() {
sed -n "s:.*<$1>\(.*\)</$1>.*:\1:p" "$root/eng/Versions.props"
}
base_property() {
git -C "$root" show "$base:eng/Versions.props" \
| sed -n "s:.*<$1>\(.*\)</$1>.*:\1:p"
}
changed() {
git -C "$root" diff --name-only "$base"...HEAD -- "$@" | grep -q .
}
require_increase() {
local property="$1"
local description="$2"
shift 2
if changed "$@"; then
local before after
before="$(base_property "$property")"
after="$(current_property "$property")"
if [[ ! "$before" =~ ^[0-9]+$ || ! "$after" =~ ^[0-9]+$ || "$after" -le "$before" ]]; then
echo "$description changed without increasing $property ($before -> $after)." >&2
exit 1
fi
fi
}
require_increase RendezvousMajorVersion "Published .NET API snapshot" \
'tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/client-public-api.txt' \
'tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/contracts-public-api.txt'
require_increase HttpContractVersion "HTTP/OpenAPI contract evidence" \
'docs/api/*.json' \
'tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/*.json' \
':(exclude)tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/connection-ticket.json'
require_increase UdpContractVersion "UDP contract evidence" \
'tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/*.hex'
require_increase ConnectionTicketFormatVersion "Connection-ticket format evidence" \
'tests/FinalFactory.Rendezvous.Tests/TestData/Contracts/v*/connection-ticket.json'
echo "Compatibility changes are paired with the required version increase."
+27
View File
@@ -0,0 +1,27 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
tag="${1:-${GITHUB_REF_NAME:-}}"
version="$(sed -n 's:.*<RendezvousVersion>\(.*\)</RendezvousVersion>.*:\1:p' "$root/eng/Versions.props")"
if [[ "$tag" != "v$version" ]]; then
echo "Release tag $tag does not match central version v$version." >&2
exit 1
fi
if [[ -n "$(git -C "$root" status --porcelain --untracked-files=normal)" ]]; then
echo "Release tag checkout is not clean." >&2
exit 1
fi
if [[ "$(git -C "$root" tag --points-at HEAD --list "$tag")" != "$tag" ]]; then
echo "Release tag $tag does not point at the checked-out commit." >&2
exit 1
fi
if ! grep -Eq "^## $version - [0-9]{4}-[0-9]{2}-[0-9]{2}$" "$root/CHANGELOG.md"; then
echo "CHANGELOG.md must contain a dated heading for $version." >&2
exit 1
fi
if grep -Eq "^## $version - Unreleased$" "$root/CHANGELOG.md"; then
echo "Release $version is still marked Unreleased." >&2
exit 1
fi
+26
View File
@@ -0,0 +1,26 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:?usage: finalize-release-candidate.sh VERSION RELEASE_DIRECTORY}"
release_dir="${2:?usage: finalize-release-candidate.sh VERSION RELEASE_DIRECTORY}"
container_sbom="$release_dir/FinalFactory.Rendezvous.Container.$version.spdx.json"
[[ -s "$container_sbom" ]] || {
echo "Container SBOM is missing or empty: $container_sbom" >&2
exit 1
}
(
cd "$release_dir"
find . -maxdepth 1 -type f ! -name checksums.sha256 -printf '%f\n' \
| LC_ALL=C sort \
| xargs sha256sum >checksums.sha256
)
python3 "$root/eng/release_artifacts.py" verify \
--root "$root" \
--release-dir "$release_dir" \
--version "$version" \
--phase publish-ready
echo "Finalized publish-ready release candidate $version"
+23
View File
@@ -0,0 +1,23 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:?usage: finalize-signing-ready-release.sh VERSION RELEASE_DIRECTORY}"
release_dir="${2:?usage: finalize-signing-ready-release.sh VERSION RELEASE_DIRECTORY}"
(
cd "$release_dir"
find . -maxdepth 1 -type f \
! -name checksums.sha256 \
! -name checksums.sha256.bundle \
-printf '%f\n' \
| LC_ALL=C sort \
| xargs sha256sum >checksums.sha256
)
python3 "$root/eng/release_artifacts.py" verify \
--root "$root" \
--release-dir "$release_dir" \
--version "$version" \
--phase signing-ready
echo "Finalized signing-ready release $version"
+153
View File
@@ -0,0 +1,153 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:?usage: publish-release.sh VERSION RELEASE_DIRECTORY}"
release_dir="${2:?usage: publish-release.sh VERSION RELEASE_DIRECTORY}"
api="${RENDEZVOUS_GITEA_API:-https://git.finalfactory.de/api/v1}"
registry="${RENDEZVOUS_CONTAINER_REGISTRY:-git.finalfactory.de}"
image="$registry/heikyu/rendezvous:$version"
token="${RENDEZVOUS_RELEASE_TOKEN:?RENDEZVOUS_RELEASE_TOKEN is required}"
username="${RENDEZVOUS_RELEASE_USERNAME:?RENDEZVOUS_RELEASE_USERNAME is required}"
release_builder="${RENDEZVOUS_RELEASE_BUILDER:?RENDEZVOUS_RELEASE_BUILDER is required}"
docker_config="$(mktemp -d)"
curl_config="$(mktemp)"
release_request=""
release_response=""
cleanup() {
[[ -z "$release_request" ]] || rm -f "$release_request"
[[ -z "$release_response" ]] || rm -f "$release_response"
rm -f "$curl_config"
rm -rf "$docker_config"
}
trap cleanup EXIT
chmod 0700 "$docker_config"
chmod 0600 "$curl_config"
printf 'header = "Authorization: token %s"\n' "$token" >"$curl_config"
export DOCKER_CONFIG="$docker_config"
for command in cosign curl docker dotnet jq; do
command -v "$command" >/dev/null || {
echo "Required publication command is unavailable: $command" >&2
exit 1
}
done
run_release_builder() {
docker run --rm \
--user "$(id -u):$(id -g)" \
--volume "$root:/source:ro" \
--volume "$release_dir:$release_dir" \
--workdir /source \
"$release_builder" "$@"
}
"$root/scripts/check-release-tag.sh" "v$version"
"$root/scripts/verify-release.sh" "$version" "$release_dir" publish-ready
require_absent() {
local description="$1"
local url="$2"
local status
status="$(curl --silent --show-error --output /dev/null --write-out '%{http_code}' \
--config "$curl_config" "$url")"
if [[ "$status" != 404 ]]; then
echo "$description must not exist before publication (HTTP $status)." >&2
exit 1
fi
}
# Gitea package versions are immutable. Require all destinations to be empty
# before the first write so a tag can never become a silent partial rerun.
require_absent "Client package $version" \
"$api/packages/HeiKyu/nuget/FinalFactory.Rendezvous.Client/$version"
require_absent "Contracts package $version" \
"$api/packages/HeiKyu/nuget/FinalFactory.Rendezvous.Contracts/$version"
require_absent "Container $version" \
"$api/packages/HeiKyu/container/rendezvous/$version"
require_absent "Release v$version" \
"$api/repos/HeiKyu/Rendezvous/releases/tags/v$version"
feed="https://git.finalfactory.de/api/packages/HeiKyu/nuget/index.json"
for package in \
"$release_dir/FinalFactory.Rendezvous.Contracts.$version.nupkg" \
"$release_dir/FinalFactory.Rendezvous.Client.$version.nupkg"; do
dotnet nuget push "$package" \
--source "$feed" \
--api-key "$token" \
--timeout 300
done
printf '%s' "$token" | docker login "$registry" --username "$username" --password-stdin
expected_image_id="$(jq -er '.containerImageId' "$release_dir/release-provenance.json")"
current_image_id="$(docker image inspect --format '{{.Id}}' "$image")"
if [[ "$current_image_id" != "$expected_image_id" ]]; then
echo "Local release tag changed after staging ($expected_image_id -> $current_image_id)." >&2
exit 1
fi
docker push "$image"
digest_ref="$(docker inspect --format '{{index .RepoDigests 0}}' "$image")"
if [[ ! "$digest_ref" =~ ^git\.finalfactory\.de/heikyu/rendezvous@sha256:[0-9a-f]{64}$ ]]; then
echo "Registry did not return an immutable Rendezvous image digest: $digest_ref" >&2
exit 1
fi
run_release_builder python3 eng/release_artifacts.py record-container-digest \
--release-dir "$release_dir" \
--digest "$digest_ref"
cosign public-key --key env://COSIGN_PRIVATE_KEY >"$release_dir/cosign.pub"
run_release_builder ./scripts/finalize-signing-ready-release.sh \
"$version" "$release_dir"
cosign sign --yes --key env://COSIGN_PRIVATE_KEY "$digest_ref"
cosign attest --yes \
--key env://COSIGN_PRIVATE_KEY \
--type https://finalfactory.de/rendezvous/release-provenance/v1 \
--predicate "$release_dir/release-provenance.json" \
"$digest_ref"
cosign sign-blob --yes \
--key env://COSIGN_PRIVATE_KEY \
--bundle "$release_dir/checksums.sha256.bundle" \
"$release_dir/checksums.sha256"
"$root/scripts/verify-release.sh" "$version" "$release_dir" published
cosign verify --key "$release_dir/cosign.pub" "$digest_ref" >/dev/null
cosign verify-attestation \
--key "$release_dir/cosign.pub" \
--type https://finalfactory.de/rendezvous/release-provenance/v1 \
"$digest_ref" >/dev/null
cosign verify-blob \
--key "$release_dir/cosign.pub" \
--bundle "$release_dir/checksums.sha256.bundle" \
"$release_dir/checksums.sha256" >/dev/null
release_request="$(mktemp)"
release_response="$(mktemp)"
prerelease=false
if [[ "$version" == *-* ]]; then
prerelease=true
fi
jq -n \
--arg tag "v$version" \
--arg commit "${GITHUB_SHA:?GITHUB_SHA is required}" \
--arg digest "$digest_ref" \
--argjson prerelease "$prerelease" \
--rawfile changelog "$release_dir/CHANGELOG.md" \
'{tag_name:$tag,target_commitish:$commit,name:("Rendezvous " + $tag),body:($changelog + "\n\n## Immutable container\n\n`" + $digest + "`\n"),draft:false,prerelease:$prerelease}' \
>"$release_request"
curl --fail --silent --show-error \
--request POST \
--config "$curl_config" \
--header 'Content-Type: application/json' \
--data-binary "@$release_request" \
"$api/repos/HeiKyu/Rendezvous/releases" >"$release_response"
release_id="$(jq -er '.id' "$release_response")"
for artifact in "$release_dir"/*; do
curl --fail --silent --show-error \
--request POST \
--config "$curl_config" \
--form "attachment=@$artifact" \
"$api/repos/HeiKyu/Rendezvous/releases/$release_id/assets?name=$(basename "$artifact")" \
>/dev/null
done
echo "Published immutable release v$version with container $digest_ref"
+58
View File
@@ -0,0 +1,58 @@
#!/usr/bin/env bash
set -euo pipefail
ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
PROFILE="${RENDEZVOUS_CAPACITY_PROFILE:-quick}"
OUTPUT="${RENDEZVOUS_CAPACITY_OUTPUT:-$ROOT/artifacts/capacity/rendezvous-capacity-v2.json}"
CPUSET="${RENDEZVOUS_CAPACITY_CPUSET:-}"
PROJECT="$ROOT/tests/FinalFactory.Rendezvous.Capacity/FinalFactory.Rendezvous.Capacity.csproj"
TESTS="$ROOT/tests/FinalFactory.Rendezvous.Tests/FinalFactory.Rendezvous.Tests.csproj"
if [[ "$PROFILE" != quick && "$PROFILE" != candidate ]]; then
printf 'RENDEZVOUS_CAPACITY_PROFILE must be quick or candidate.\n' >&2
exit 2
fi
command -v dotnet >/dev/null || {
printf 'Missing required command: dotnet\n' >&2
exit 2
}
if [[ -n "$CPUSET" ]]; then
command -v taskset >/dev/null || {
printf 'taskset is required when RENDEZVOUS_CAPACITY_CPUSET is set.\n' >&2
exit 2
}
fi
cd "$ROOT"
export RENDEZVOUS_EVIDENCE_COMMIT="$(git rev-parse HEAD)"
if [[ -n "$(git status --porcelain)" ]]; then
export RENDEZVOUS_EVIDENCE_TREE_STATE=dirty
else
export RENDEZVOUS_EVIDENCE_TREE_STATE=clean
fi
if [[ "$PROFILE" == candidate && "$RENDEZVOUS_EVIDENCE_TREE_STATE" != clean ]]; then
printf 'Candidate evidence requires a clean source tree.\n' >&2
exit 2
fi
export RENDEZVOUS_EVIDENCE_CPUSET="${CPUSET:-unrestricted}"
export RENDEZVOUS_EVIDENCE_COMMAND="RENDEZVOUS_CAPACITY_PROFILE=$PROFILE RENDEZVOUS_CAPACITY_CPUSET=${CPUSET:-unrestricted} ./scripts/run-capacity-gate.sh"
dotnet restore "$ROOT/Rendezvous.slnx" --locked-mode
dotnet build "$ROOT/Rendezvous.slnx" --configuration Release --no-restore
filter='FullyQualifiedName~TrackerCapacityFailsClosedWithoutGrowingAndAWindowResetRecovers|FullyQualifiedName~OptionalTrafficCannotConsumeTheLeaseOperationReserve|FullyQualifiedName~ConcurrentAbusiveBurstStaysBoundedAndCannotBlockCriticalHttp|FullyQualifiedName~HttpOverloadIsTypedAndOversizedBodiesAreRejectedBeforeDispatch|FullyQualifiedName~WallClockMovementDoesNotExpireOrExtendLease|FullyQualifiedName~RepeatedMutableDeadlineRefreshesKeepOneScheduledEntryPerKey|FullyQualifiedName~RepeatedPrincipalRevocationCanExtendButCannotShortenProtection|FullyQualifiedName~RestartHasNewGenerationAndNoEphemeralState|FullyQualifiedName~RestartReturnsTypedUnavailabilityThenAllowsHostReregistration|FullyQualifiedName~DrainRejectsNewWorkAllowsInflightCompletionThenClearsState|FullyQualifiedName~UnavailableStoreFailsNewAuthorizationClosedAndErasesActiveState|FullyQualifiedName~KeyRotationHonorsOverlapAndRejectsRetiredKeys|FullyQualifiedName~OperatorSurfaceSeparatesAuthenticationConfirmsActionsAndRedactsInspection|FullyQualifiedName~SigtermDrainsThenReleasesHttpAndUdpSockets|FullyQualifiedName~ProductionTransportSoakKeepsHandlesMemoryAndSocketsBounded|FullyQualifiedName~NativeLiteNetLibRequestsIntroduceTheAuthorizedPair'
dotnet test "$TESTS" --configuration Release --no-build --filter "$filter" \
--logger 'console;verbosity=minimal'
mkdir -p "$(dirname "$OUTPUT")"
arguments=(
dotnet run --project "$PROJECT" --configuration Release --no-build --
--profile "$PROFILE" --output "$OUTPUT"
)
if [[ -n "$CPUSET" ]]; then
taskset -c "$CPUSET" "${arguments[@]}"
else
"${arguments[@]}"
fi
printf 'Capacity and resilience gate passed; evidence: %s\n' "$OUTPUT"
+148
View File
@@ -0,0 +1,148 @@
#!/usr/bin/env bash
set -euo pipefail
ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
SERVICE_URL="${RENDEZVOUS_SMOKE_HTTP_URL:-http://127.0.0.1:8080/}"
MEDIATOR="${RENDEZVOUS_SMOKE_UDP_ENDPOINT:-127.0.0.1:9050}"
TIMEOUT_SECONDS="${RENDEZVOUS_SMOKE_TIMEOUT_SECONDS:-30}"
LOCAL_KEY="${RENDEZVOUS_SMOKE_LOCAL_KEY:-$ROOT/deploy/compose/secrets/signing-key}"
PROJECT="$ROOT/src/FinalFactory.Rendezvous.TestClient/FinalFactory.Rendezvous.TestClient.csproj"
BUILD_CONFIGURATION="${RENDEZVOUS_SMOKE_CONFIGURATION:-Release}"
GAME_ID="${RENDEZVOUS_SMOKE_GAME_ID:-space-game}"
ENVIRONMENT_ID="${RENDEZVOUS_SMOKE_ENVIRONMENT_ID:-smoke}"
REGION="${RENDEZVOUS_SMOKE_REGION:-local}"
PROTOCOL_VERSION="${RENDEZVOUS_SMOKE_PROTOCOL_VERSION:-1}"
for command in curl date dotnet jq mktemp od openssl tail tr wc; do
command -v "$command" >/dev/null || {
printf 'Missing required command: %s\n' "$command" >&2
exit 2
}
done
if [[ ! "$TIMEOUT_SECONDS" =~ ^[0-9]+$ ]] || (( TIMEOUT_SECONDS < 1 || TIMEOUT_SECONDS > 300 )); then
printf 'RENDEZVOUS_SMOKE_TIMEOUT_SECONDS must be an integer from 1 through 300.\n' >&2
exit 2
fi
if [[ ! "$PROTOCOL_VERSION" =~ ^[0-9]+$ ]] || (( PROTOCOL_VERSION < 1 )); then
printf 'RENDEZVOUS_SMOKE_PROTOCOL_VERSION must be a positive integer.\n' >&2
exit 2
fi
for scoped_value in "$GAME_ID" "$ENVIRONMENT_ID" "$REGION"; do
if [[ -z "$scoped_value" ]]; then
printf 'Smoke game, environment, and region values must not be empty.\n' >&2
exit 2
fi
done
base64url() {
openssl base64 -A | tr '+/' '-_' | tr -d '='
}
local_credential() {
if [[ ! -f "$LOCAL_KEY" ]] || [[ "$(wc -c < "$LOCAL_KEY")" -ne 32 ]]; then
printf 'Local Compose smoke key must be exactly 32 bytes: %s\n' "$LOCAL_KEY" >&2
exit 2
fi
local now expires nonce payload encoded signed hex signature
now="$(date +%s)"
expires="$((now + 600))"
nonce="$(openssl rand -hex 16)"
payload="$(jq -cn \
--arg issuer final-factory-rendezvous-smoke \
--arg audience rendezvous-service \
--arg subject local-smoke-host \
--arg kind dedicatedPublisher \
--arg gameId "$GAME_ID" \
--arg environmentId "$ENVIRONMENT_ID" \
--arg region "$REGION" \
--arg nonce "$nonce" \
--argjson now "$now" \
--argjson expires "$expires" \
'{version:1,issuer:$issuer,audience:$audience,subject:$subject,kind:$kind,gameId:$gameId,environmentId:$environmentId,regions:[$region],permissions:[],issuedAtUnixSeconds:$now,notBeforeUnixSeconds:$now,expiresAtUnixSeconds:$expires,nonce:$nonce}')"
encoded="$(printf '%s' "$payload" | base64url)"
signed="rv1.local-smoke-1.$encoded"
hex="$(od -An -v -tx1 "$LOCAL_KEY" | tr -d ' \n')"
signature="$(printf '%s' "$signed" \
| openssl dgst -sha256 -mac HMAC -macopt "hexkey:$hex" -binary \
| base64url)"
printf '%s.%s' "$signed" "$signature"
}
credential="${RENDEZVOUS_PUBLISHER_CREDENTIAL:-}"
if [[ -z "$credential" ]]; then
credential="$(local_credential)"
fi
export RENDEZVOUS_PUBLISHER_CREDENTIAL="$credential"
curl --fail --silent --show-error --max-time 5 "${SERVICE_URL%/}/health/live" >/dev/null
curl --fail --silent --show-error --max-time 5 "${SERVICE_URL%/}/health/ready" >/dev/null
temp_dir="$(mktemp -d)"
host_log="$temp_dir/host.jsonl"
join_log="$temp_dir/join.jsonl"
host_pid=''
cleanup() {
local status="$?"
if [[ -n "$host_pid" ]] && kill -0 "$host_pid" 2>/dev/null; then
kill -TERM "$host_pid" 2>/dev/null || true
wait "$host_pid" 2>/dev/null || true
fi
if [[ "$status" -ne 0 ]]; then
printf 'Deployment smoke failed; sanitized diagnostic events follow.\n' >&2
[[ -f "$host_log" ]] && jq -c . "$host_log" >&2 || true
[[ -f "$join_log" ]] && jq -c . "$join_log" >&2 || true
fi
rm -rf "$temp_dir"
return "$status"
}
trap cleanup EXIT
trap 'exit 130' INT
trap 'exit 143' TERM
dotnet run --project "$PROJECT" --configuration "$BUILD_CONFIGURATION" --no-build -- \
host --service "$SERVICE_URL" --mediator "$MEDIATOR" \
--game "$GAME_ID" --environment "$ENVIRONMENT_ID" --region "$REGION" --protocol "$PROTOCOL_VERSION" \
--script --json --exit-after-echo --timeout-seconds "$TIMEOUT_SECONDS" \
>"$host_log" 2>&1 &
host_pid="$!"
ready=false
for ((iteration = 0; iteration < TIMEOUT_SECONDS * 4; iteration++)); do
if jq -e 'select(.event == "host.ready")' "$host_log" >/dev/null 2>&1; then
ready=true
break
fi
if ! kill -0 "$host_pid" 2>/dev/null; then
printf 'Host diagnostic stopped before it became ready.\n' >&2
jq -c . "$host_log" >&2 || true
exit 1
fi
sleep 0.25
done
if [[ "$ready" != true ]]; then
printf 'Host diagnostic did not become ready within %s seconds.\n' "$TIMEOUT_SECONDS" >&2
exit 1
fi
listing_id="$(jq -r 'select(.event == "host.registered") | .listingId' "$host_log" | tail -n 1)"
if [[ -z "$listing_id" || "$listing_id" == null ]]; then
printf 'Host diagnostic did not report a listing ID.\n' >&2
exit 1
fi
dotnet run --project "$PROJECT" --configuration "$BUILD_CONFIGURATION" --no-build -- \
join --service "$SERVICE_URL" --mediator "$MEDIATOR" \
--game "$GAME_ID" --environment "$ENVIRONMENT_ID" --region "$REGION" --protocol "$PROTOCOL_VERSION" \
--listing "$listing_id" --script --json --timeout-seconds "$TIMEOUT_SECONDS" \
>"$join_log" 2>&1
wait "$host_pid"
host_pid=''
jq -e 'select(.event == "host.direct-traffic" and .status == "verified")' "$host_log" >/dev/null
jq -e 'select(.event == "host.deregistered" and .status == "complete")' "$host_log" >/dev/null
jq -e 'select(.event == "join.direct-traffic" and .status == "verified")' "$join_log" >/dev/null
jq -e 'select(.event == "join.outcome-report" and .status == "accepted")' "$join_log" >/dev/null
printf 'Rendezvous deployment smoke passed: HTTP live/ready and authenticated UDP mediation/direct traffic.\n'
+78
View File
@@ -0,0 +1,78 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:?usage: verify-real-consumers.sh VERSION RELEASE_DIRECTORY}"
release_dir="${2:?usage: verify-real-consumers.sh VERSION RELEASE_DIRECTORY}"
manifest="$root/eng/consumer-revisions.json"
work="$(mktemp -d "${TMPDIR:-/tmp}/rendezvous-consumers.XXXXXX")"
cleanup() {
rm -rf "$work"
}
trap cleanup EXIT
for command in dotnet git jq python3; do
command -v "$command" >/dev/null || {
echo "Required consumer verification command is unavailable: $command" >&2
exit 1
}
done
python3 "$root/eng/release_artifacts.py" consumer-config \
--local-source "$release_dir" \
--output "$work/NuGet.config"
count="$(jq '.consumers | length' "$manifest")"
for ((index = 0; index < count; index++)); do
name="$(jq -r ".consumers[$index].name" "$manifest")"
repository="$(jq -r ".consumers[$index].repository" "$manifest")"
revision="$(jq -r ".consumers[$index].revision" "$manifest")"
project_relative="$(jq -r ".consumers[$index].project" "$manifest")"
checkout="$work/$name"
git -c init.defaultBranch=main init --quiet "$checkout"
git -C "$checkout" remote add origin "$repository"
git -C "$checkout" fetch --quiet --depth 1 origin "$revision"
GIT_LFS_SKIP_SMUDGE=1 git -C "$checkout" checkout --quiet --detach FETCH_HEAD
[[ "$(git -C "$checkout" rev-parse HEAD)" == "$revision" ]] || {
echo "$name did not resolve the pinned consumer revision." >&2
exit 1
}
project="$checkout/$project_relative"
[[ -f "$project" ]] || {
echo "$name consumer project does not exist at $project_relative." >&2
exit 1
}
targets="$work/$name.Rendezvous.Consumer.targets"
cat >"$targets" <<EOF
<Project>
<ItemGroup Condition="'\$(MSBuildProjectFullPath)' == '$project'">
<PackageReference Include="FinalFactory.Rendezvous.Client" Version="[$version]" />
<PackageReference Include="FinalFactory.Rendezvous.Contracts" Version="[$version]" />
</ItemGroup>
</Project>
EOF
packages="$work/packages-$name"
dotnet restore "$project" \
-p:CustomAfterMicrosoftCommonTargets="$targets" \
-p:RestorePackagesWithLockFile=false \
-p:RestoreLockedMode=false \
--packages "$packages" \
--configfile "$work/NuGet.config" \
--force-evaluate
assets=""
while IFS= read -r candidate_assets; do
if grep -Fq "FinalFactory.Rendezvous.Client/$version" "$candidate_assets"; then
assets="$candidate_assets"
break
fi
done < <(find "$checkout" -path '*/obj/project.assets.json' -type f -print)
[[ -n "$assets" ]] || {
echo "$name restore did not produce assets for the injected Rendezvous references." >&2
exit 1
}
python3 "$root/eng/release_artifacts.py" consumer \
--assets "$assets" \
--version "$version"
echo "Verified $name at $revision can pin and restore Rendezvous $version."
done
+13
View File
@@ -0,0 +1,13 @@
#!/usr/bin/env bash
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
version="${1:?usage: verify-release.sh VERSION [RELEASE_DIRECTORY] [PHASE]}"
release_dir="${2:-$root/artifacts/release/$version}"
phase="${3:-build}"
python3 "$root/eng/release_artifacts.py" verify \
--root "$root" \
--release-dir "$release_dir" \
--version "$version" \
--phase "$phase"
@@ -0,0 +1,183 @@
using FinalFactory.Rendezvous.Contracts;
using LiteNetLib;
namespace FinalFactory.Rendezvous.Client;
public enum RendezvousConnectionOutcomeSource
{
RendezvousService = 1,
LocalTraversal = 2,
RemoteHost = 3,
Caller = 4,
Lifecycle = 5,
}
public enum RendezvousConnectionFailureCategory
{
None = 0,
Directory = 1,
Compatibility = 2,
Authorization = 3,
Capacity = 4,
HostPresence = 5,
Service = 6,
Mediation = 7,
NatTraversal = 8,
DirectConnection = 9,
Lifecycle = 10,
}
public enum RendezvousConnectionPhase
{
Directory = 1,
Authorization = 2,
Mediation = 3,
NatTraversal = 4,
DirectConnection = 5,
Complete = 6,
}
public sealed class RendezvousConnectionOutcome
{
private readonly NetworkEndpoint? _dedicatedFallback;
private RendezvousConnectionOutcome(
ConnectionOutcomeKind kind,
RendezvousConnectionOutcomeSource source,
RendezvousConnectionFailureCategory category,
RendezvousConnectionPhase phase,
TimeSpan elapsed,
RendezvousErrorCode? serviceError,
NetworkEndpoint? dedicatedFallback,
NetPeer? peer)
{
if (elapsed < TimeSpan.Zero)
{
throw new ArgumentOutOfRangeException(nameof(elapsed));
}
if (dedicatedFallback is not null
&& !ContractValidation.IsNetworkEndpointValid(dedicatedFallback))
{
throw new ArgumentException("The dedicated fallback endpoint is invalid.", nameof(dedicatedFallback));
}
Kind = kind;
Source = source;
Category = category;
Phase = phase;
Elapsed = elapsed;
ServiceError = serviceError;
_dedicatedFallback = RendezvousEndpoint.Copy(dedicatedFallback);
Peer = peer;
}
public ConnectionOutcomeKind Kind { get; }
public RendezvousConnectionOutcomeSource Source { get; }
public RendezvousConnectionFailureCategory Category { get; }
public RendezvousConnectionPhase Phase { get; }
public TimeSpan Elapsed { get; }
public RendezvousErrorCode? ServiceError { get; }
public NetworkEndpoint? DedicatedFallback => RendezvousEndpoint.Copy(_dedicatedFallback);
public NetPeer? Peer { get; }
public bool IsSuccess => Kind == ConnectionOutcomeKind.Connected;
public bool HasDedicatedFallback => _dedicatedFallback is not null;
public static RendezvousConnectionOutcome FromServiceError(
RendezvousErrorCode error,
TimeSpan elapsed,
NetworkEndpoint? dedicatedFallback = null)
{
if (error == RendezvousErrorCode.None)
{
throw new ArgumentException("A service failure outcome requires an error.", nameof(error));
}
(ConnectionOutcomeKind kind, RendezvousConnectionFailureCategory category, RendezvousConnectionPhase phase) =
error switch
{
RendezvousErrorCode.NotFound => (
ConnectionOutcomeKind.DirectoryNotFound,
RendezvousConnectionFailureCategory.Directory,
RendezvousConnectionPhase.Directory),
RendezvousErrorCode.Expired => (
ConnectionOutcomeKind.AttemptExpired,
RendezvousConnectionFailureCategory.Authorization,
RendezvousConnectionPhase.Authorization),
RendezvousErrorCode.IncompatibleProtocol => (
ConnectionOutcomeKind.IncompatibleProtocol,
RendezvousConnectionFailureCategory.Compatibility,
RendezvousConnectionPhase.Directory),
RendezvousErrorCode.AuthenticationRequired
or RendezvousErrorCode.Forbidden
or RendezvousErrorCode.ReplayRejected => (
ConnectionOutcomeKind.Unauthorized,
RendezvousConnectionFailureCategory.Authorization,
RendezvousConnectionPhase.Authorization),
RendezvousErrorCode.RateLimited
or RendezvousErrorCode.CapacityExceeded => (
ConnectionOutcomeKind.RateLimited,
RendezvousConnectionFailureCategory.Capacity,
RendezvousConnectionPhase.Authorization),
RendezvousErrorCode.StaleHost => (
ConnectionOutcomeKind.NoHostPresence,
RendezvousConnectionFailureCategory.HostPresence,
RendezvousConnectionPhase.Mediation),
RendezvousErrorCode.ServiceUnavailable => (
ConnectionOutcomeKind.ServiceUnavailable,
RendezvousConnectionFailureCategory.Service,
RendezvousConnectionPhase.Authorization),
_ => (
ConnectionOutcomeKind.ServiceRejected,
RendezvousConnectionFailureCategory.Service,
RendezvousConnectionPhase.Authorization),
};
return new(
kind,
RendezvousConnectionOutcomeSource.RendezvousService,
category,
phase,
elapsed,
error,
dedicatedFallback,
null);
}
public static ConnectionElapsedBucket BucketElapsed(TimeSpan elapsed)
{
if (elapsed < TimeSpan.Zero)
{
throw new ArgumentOutOfRangeException(nameof(elapsed));
}
return elapsed.TotalSeconds switch
{
< 1 => ConnectionElapsedBucket.UnderOneSecond,
< 5 => ConnectionElapsedBucket.OneToFiveSeconds,
< 15 => ConnectionElapsedBucket.FiveToFifteenSeconds,
< 30 => ConnectionElapsedBucket.FifteenToThirtySeconds,
_ => ConnectionElapsedBucket.ThirtySecondsOrMore,
};
}
public override string ToString() =>
$"[RendezvousConnectionOutcome {Kind}; {Source}; credentials redacted]";
internal static RendezvousConnectionOutcome Create(
ConnectionOutcomeKind kind,
RendezvousConnectionOutcomeSource source,
RendezvousConnectionFailureCategory category,
RendezvousConnectionPhase phase,
TimeSpan elapsed,
NetworkEndpoint? dedicatedFallback = null,
NetPeer? peer = null) => new(
kind,
source,
category,
phase,
elapsed,
null,
dedicatedFallback,
peer);
}
@@ -0,0 +1,35 @@
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Client;
public sealed class RendezvousConnectionStartResult
{
internal RendezvousConnectionStartResult(
CreateJoinAttemptResponse? attempt,
RendezvousConnectionOutcome? outcome)
{
if ((attempt is null) == (outcome is null))
{
throw new ArgumentException(
"A connection start result requires exactly one attempt or terminal outcome.");
}
Attempt = attempt;
Outcome = outcome;
}
public CreateJoinAttemptResponse? Attempt { get; }
public RendezvousConnectionOutcome? Outcome { get; }
public bool IsReadyForTraversal => Attempt is not null;
public bool IsCompleted => Outcome is not null;
public static RendezvousConnectionStartResult ReadyForTraversal(
CreateJoinAttemptResponse attempt) => new(
attempt ?? throw new ArgumentNullException(nameof(attempt)),
null);
public static RendezvousConnectionStartResult Completed(
RendezvousConnectionOutcome outcome) => new(
null,
outcome ?? throw new ArgumentNullException(nameof(outcome)));
}
@@ -0,0 +1,232 @@
using System.Security.Cryptography;
using System.Text;
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Client;
public enum ConnectionTicketConsumptionResult
{
Accepted = 1,
NotFound = 2,
Expired = 3,
Rejected = 4,
AlreadyConsumed = 5,
Revoked = 6,
}
public sealed class ConnectionTicketValidator : IDisposable
{
private readonly object _gate = new();
private readonly Dictionary<JoinAttemptId, TicketEntry> _tickets = [];
private readonly int _maximumAuthorizedTickets;
private readonly IConnectionTicketClock _clock;
private readonly byte[] _fingerprintKey = new byte[32];
private bool _disposed;
public ConnectionTicketValidator(int maximumAuthorizedTickets = 1_024)
: this(maximumAuthorizedTickets, new SystemConnectionTicketClock())
{
}
internal ConnectionTicketValidator(
int maximumAuthorizedTickets,
IConnectionTicketClock clock)
{
if (maximumAuthorizedTickets is < 1 or > 10_000)
{
throw new ArgumentOutOfRangeException(nameof(maximumAuthorizedTickets));
}
_maximumAuthorizedTickets = maximumAuthorizedTickets;
_clock = clock ?? throw new ArgumentNullException(nameof(clock));
RandomNumberGenerator.Fill(_fingerprintKey);
}
public bool TryAuthorize(
JoinAttemptId attemptId,
string connectionTicket,
DateTimeOffset expiresAt)
{
lock (_gate)
{
ThrowIfDisposed();
DateTimeOffset now = _clock.UtcNow;
if (attemptId.Value == Guid.Empty
|| !ContractValidation.IsConnectionTicketValid(connectionTicket)
|| expiresAt <= now)
{
return false;
}
RemoveExpired(now);
byte[] fingerprint = Fingerprint(connectionTicket);
if (_tickets.TryGetValue(attemptId, out TicketEntry? current))
{
bool idempotent = current.State == TicketState.Active
&& current.ExpiresAt == expiresAt
&& CryptographicOperations.FixedTimeEquals(current.Fingerprint, fingerprint);
CryptographicOperations.ZeroMemory(fingerprint);
return idempotent;
}
if (_tickets.Count >= _maximumAuthorizedTickets)
{
CryptographicOperations.ZeroMemory(fingerprint);
return false;
}
_tickets.Add(attemptId, new(fingerprint, expiresAt));
return true;
}
}
public ConnectionTicketConsumptionResult Consume(
JoinAttemptId attemptId,
string connectionTicket)
{
lock (_gate)
{
ThrowIfDisposed();
DateTimeOffset now = _clock.UtcNow;
if (attemptId.Value == Guid.Empty
|| !ContractValidation.IsConnectionTicketValid(connectionTicket))
{
return ConnectionTicketConsumptionResult.Rejected;
}
if (!_tickets.TryGetValue(attemptId, out TicketEntry? entry))
{
RemoveExpired(now);
return ConnectionTicketConsumptionResult.NotFound;
}
if (entry.ExpiresAt <= now)
{
Remove(attemptId, entry);
return ConnectionTicketConsumptionResult.Expired;
}
if (entry.State == TicketState.Revoked)
{
return ConnectionTicketConsumptionResult.Revoked;
}
if (entry.State == TicketState.Consumed)
{
return ConnectionTicketConsumptionResult.AlreadyConsumed;
}
byte[] supplied = Fingerprint(connectionTicket);
bool matches = CryptographicOperations.FixedTimeEquals(entry.Fingerprint, supplied);
CryptographicOperations.ZeroMemory(supplied);
if (!matches)
{
return ConnectionTicketConsumptionResult.Rejected;
}
entry.State = TicketState.Consumed;
return ConnectionTicketConsumptionResult.Accepted;
}
}
public bool Revoke(JoinAttemptId attemptId)
{
lock (_gate)
{
ThrowIfDisposed();
RemoveExpired(_clock.UtcNow);
if (!_tickets.TryGetValue(attemptId, out TicketEntry? entry))
{
return false;
}
entry.State = TicketState.Revoked;
CryptographicOperations.ZeroMemory(entry.Fingerprint);
return true;
}
}
public void Dispose()
{
lock (_gate)
{
if (_disposed)
{
return;
}
foreach (TicketEntry entry in _tickets.Values)
{
CryptographicOperations.ZeroMemory(entry.Fingerprint);
}
_tickets.Clear();
CryptographicOperations.ZeroMemory(_fingerprintKey);
_disposed = true;
}
}
public override string ToString() => "[ConnectionTicketValidator: tickets and key redacted]";
private byte[] Fingerprint(string ticket)
{
byte[] encoded = Encoding.ASCII.GetBytes(ticket);
try
{
using HMACSHA256 hmac = new(_fingerprintKey);
return hmac.ComputeHash(encoded);
}
finally
{
CryptographicOperations.ZeroMemory(encoded);
}
}
private void RemoveExpired(DateTimeOffset now)
{
foreach (KeyValuePair<JoinAttemptId, TicketEntry> item in _tickets
.Where(item => item.Value.ExpiresAt <= now)
.ToArray())
{
Remove(item.Key, item.Value);
}
}
private void Remove(JoinAttemptId attemptId, TicketEntry entry)
{
CryptographicOperations.ZeroMemory(entry.Fingerprint);
_tickets.Remove(attemptId);
}
private void ThrowIfDisposed()
{
if (_disposed)
{
throw new ObjectDisposedException(nameof(ConnectionTicketValidator));
}
}
private sealed class TicketEntry(byte[] fingerprint, DateTimeOffset expiresAt)
{
public byte[] Fingerprint { get; } = fingerprint;
public DateTimeOffset ExpiresAt { get; } = expiresAt;
public TicketState State { get; set; }
}
private enum TicketState
{
Active = 0,
Consumed = 1,
Revoked = 2,
}
}
internal interface IConnectionTicketClock
{
DateTimeOffset UtcNow { get; }
}
internal sealed class SystemConnectionTicketClock : IConnectionTicketClock
{
public DateTimeOffset UtcNow => DateTimeOffset.UtcNow;
}
@@ -7,10 +7,12 @@
<PackageId>FinalFactory.Rendezvous.Client</PackageId> <PackageId>FinalFactory.Rendezvous.Client</PackageId>
<PackageReadmeFile>README.md</PackageReadmeFile> <PackageReadmeFile>README.md</PackageReadmeFile>
<Description>Godot-independent client SDK for Final Factory Rendezvous.</Description> <Description>Godot-independent client SDK for Final Factory Rendezvous.</Description>
<PackageTags>final-factory;multiplayer;nat;godot;litenetlib</PackageTags>
</PropertyGroup> </PropertyGroup>
<ItemGroup> <ItemGroup>
<ProjectReference Include="../FinalFactory.Rendezvous.Contracts/FinalFactory.Rendezvous.Contracts.csproj" /> <ProjectReference Include="../FinalFactory.Rendezvous.Contracts/FinalFactory.Rendezvous.Contracts.csproj" />
<PackageReference Include="LiteNetLib" /> <PackageReference Include="LiteNetLib" />
<None Update="README.md" Pack="true" PackagePath="\" /> <None Update="README.md" Pack="true" PackagePath="\" />
<None Include="../../CHANGELOG.md" Pack="true" PackagePath="\" Link="CHANGELOG.md" />
</ItemGroup> </ItemGroup>
</Project> </Project>
@@ -0,0 +1,236 @@
using System.Diagnostics;
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Client;
public sealed class RendezvousJoinClient : IRendezvousJoinClient
{
private const string LeaseTokenHeader = "X-Rendezvous-Lease-Token";
private const string ClientPunchCapabilityHeader = "X-Rendezvous-Client-Punch-Capability";
private readonly RendezvousHttpTransport _transport;
public RendezvousJoinClient(
HttpClient httpClient,
RendezvousClientOptions? options = null,
IRendezvousDelay? delay = null)
{
_transport = new(httpClient, options, delay);
}
public Task<RendezvousClientResult<CreateJoinAttemptResponse>> CreateAsync(
CreateJoinAttemptRequest request,
CancellationToken cancellationToken = default)
{
if (request is null)
{
throw new ArgumentNullException(nameof(request));
}
CreateJoinAttemptRequest body = new()
{
ContractVersion = request.ContractVersion,
IdempotencyKey = request.IdempotencyKey,
GameId = request.GameId,
EnvironmentId = request.EnvironmentId,
ListingId = request.ListingId,
ProtocolVersion = request.ProtocolVersion,
};
return _transport.SendSafeAsync<CreateJoinAttemptResponse>(
() => RendezvousHttpTransport.JsonRequest(HttpMethod.Post, "v1/join-attempts", body),
cancellationToken);
}
public async Task<RendezvousConnectionStartResult> CreateConnectionAttemptAsync(
CreateJoinAttemptRequest request,
NetworkEndpoint? dedicatedFallback = null,
CancellationToken cancellationToken = default)
{
if (dedicatedFallback is not null
&& !ContractValidation.IsNetworkEndpointValid(dedicatedFallback))
{
throw new ArgumentException("The dedicated fallback endpoint is invalid.", nameof(dedicatedFallback));
}
Stopwatch elapsed = Stopwatch.StartNew();
try
{
RendezvousClientResult<CreateJoinAttemptResponse> result = await CreateAsync(
request,
cancellationToken).ConfigureAwait(false);
elapsed.Stop();
return result.IsSuccess && result.Value is not null
? RendezvousConnectionStartResult.ReadyForTraversal(result.Value)
: RendezvousConnectionStartResult.Completed(
RendezvousConnectionOutcome.FromServiceError(
result.Error,
elapsed.Elapsed,
dedicatedFallback));
}
catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
{
elapsed.Stop();
return RendezvousConnectionStartResult.Completed(
RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.Cancelled,
RendezvousConnectionOutcomeSource.Caller,
RendezvousConnectionFailureCategory.Lifecycle,
RendezvousConnectionPhase.Authorization,
elapsed.Elapsed,
dedicatedFallback));
}
}
public Task<RendezvousClientResult<bool>> CancelAsync(
CreateJoinAttemptResponse attempt,
CancellationToken cancellationToken = default)
{
if (attempt is null)
{
throw new ArgumentNullException(nameof(attempt));
}
return _transport.SendSafeAsync<bool>(
() => HeaderRequest(
HttpMethod.Delete,
$"v1/join-attempts/{attempt.AttemptId}",
ClientPunchCapabilityHeader,
RequireHeaderValue(attempt.ClientPunchCapability, nameof(attempt))),
cancellationToken);
}
public Task<RendezvousClientResult<BrowseHostJoinAttemptsResponse>> BrowseForHostAsync(
PublishedSession session,
int pageSize = ContractLimits.BrowserPageMaxItems,
string? cursor = null,
CancellationToken cancellationToken = default)
{
if (session is null)
{
throw new ArgumentNullException(nameof(session));
}
if (pageSize is < 1 or > ContractLimits.BrowserPageMaxItems)
{
throw new ArgumentOutOfRangeException(nameof(pageSize));
}
string query = $"v1/sessions/{session.ListingId}/join-attempts"
+ $"?contractVersion={ContractLimits.ContractVersion}"
+ $"&pageSize={pageSize}"
+ (cursor is null ? string.Empty : $"&cursor={Uri.EscapeDataString(cursor)}");
return _transport.SendSafeAsync<BrowseHostJoinAttemptsResponse>(
() => HeaderRequest(
HttpMethod.Get,
query,
LeaseTokenHeader,
RequireHeaderValue(session.LeaseToken, nameof(session))),
cancellationToken);
}
public async Task<RendezvousClientResult<IReadOnlyList<HostJoinAttempt>>> BrowseAllForHostAsync(
PublishedSession session,
int maximumPages = 100,
CancellationToken cancellationToken = default)
{
if (session is null)
{
throw new ArgumentNullException(nameof(session));
}
if (maximumPages is < 1 or > 1_000)
{
throw new ArgumentOutOfRangeException(nameof(maximumPages));
}
List<HostJoinAttempt> attempts = [];
string? cursor = null;
for (int page = 0; page < maximumPages; page++)
{
RendezvousClientResult<BrowseHostJoinAttemptsResponse> result =
await BrowseForHostAsync(
session,
ContractLimits.BrowserPageMaxItems,
cursor,
cancellationToken).ConfigureAwait(false);
if (!result.IsSuccess || result.Value is null)
{
return RendezvousClientResult.Failure<IReadOnlyList<HostJoinAttempt>>(
result.Error,
result.Message,
result.RetryAfterSeconds);
}
attempts.AddRange(result.Value.Items);
cursor = result.Value.NextCursor;
if (string.IsNullOrEmpty(cursor))
{
return RendezvousClientResult.Success<IReadOnlyList<HostJoinAttempt>>(
attempts.AsReadOnly());
}
}
return RendezvousClientResult.Failure<IReadOnlyList<HostJoinAttempt>>(
RendezvousErrorCode.CapacityExceeded,
$"Host invitation polling exceeded the configured {maximumPages}-page limit.");
}
public Task<RendezvousClientResult<ReportConnectionOutcomeResponse>> ReportOutcomeAsync(
CreateJoinAttemptResponse attempt,
RendezvousConnectionOutcome outcome,
CancellationToken cancellationToken = default)
{
if (attempt is null)
{
throw new ArgumentNullException(nameof(attempt));
}
if (outcome is null)
{
throw new ArgumentNullException(nameof(outcome));
}
if (!ContractValidation.IsReportableConnectionOutcome(outcome.Kind))
{
throw new ArgumentException(
"This outcome cannot be reported for an issued join attempt.",
nameof(outcome));
}
ReportConnectionOutcomeRequest body = new()
{
Outcome = outcome.Kind,
ElapsedBucket = RendezvousConnectionOutcome.BucketElapsed(outcome.Elapsed),
};
return _transport.SendSafeAsync<ReportConnectionOutcomeResponse>(
() => HeaderJsonRequest(
HttpMethod.Post,
$"v1/join-attempts/{attempt.AttemptId}/outcome",
ClientPunchCapabilityHeader,
RequireHeaderValue(attempt.ClientPunchCapability, nameof(attempt)),
body),
cancellationToken);
}
private static HttpRequestMessage HeaderRequest(
HttpMethod method,
string uri,
string header,
string value)
{
HttpRequestMessage request = new(method, uri);
request.Headers.TryAddWithoutValidation(header, value);
return request;
}
private static HttpRequestMessage HeaderJsonRequest<T>(
HttpMethod method,
string uri,
string header,
string value,
T body)
{
HttpRequestMessage request = RendezvousHttpTransport.JsonRequest(method, uri, body);
request.Headers.TryAddWithoutValidation(header, value);
return request;
}
private static string RequireHeaderValue(string value, string parameterName) =>
!string.IsNullOrWhiteSpace(value)
? value
: throw new ArgumentException("The required capability is missing.", parameterName);
}
@@ -0,0 +1,3 @@
using System.Runtime.CompilerServices;
[assembly: InternalsVisibleTo("FinalFactory.Rendezvous.Tests")]
+138 -2
View File
@@ -1,6 +1,6 @@
# FinalFactory.Rendezvous.Client # FinalFactory.Rendezvous.Client
Godot-independent .NET publisher and session-browser SDK for Rendezvous v1. Godot-independent .NET publisher, browser, join, and LiteNetLib traversal SDK for Rendezvous v1.
The package targets `netstandard2.1` and uses a caller-owned `HttpClient`. The package targets `netstandard2.1` and uses a caller-owned `HttpClient`.
```csharp ```csharp
@@ -29,6 +29,12 @@ RendezvousClientResult<PublishedSession> registered = await publisher.RegisterAs
DisplayName = "My server", DisplayName = "My server",
Visibility = ListingVisibility.Public, Visibility = ListingVisibility.Public,
Capacity = new() { CurrentPlayers = 1, MaximumPlayers = 8 }, Capacity = new() { CurrentPlayers = 1, MaximumPlayers = 8 },
DedicatedFallback = new()
{
AddressFamily = AddressFamilyKind.Ipv4,
Address = "203.0.113.40",
Port = 7777,
},
}, },
publisherCredential, publisherCredential,
cancellationToken); cancellationToken);
@@ -42,6 +48,117 @@ if (!registered.IsSuccess || registered.Value is null)
Load `publisherCredential` from the game's deployment secret boundary; never Load `publisherCredential` from the game's deployment secret boundary; never
embed it in a client build or source control. A successful registration returns a embed it in a client build or source control. A successful registration returns a
`PublishedSession` containing the lease and host-presence capabilities. `PublishedSession` containing the lease and host-presence capabilities.
Send a periodic presence request from the host's gameplay `NetManager` using the
server-controlled refresh interval and the fixed-size native token:
```csharp
string presenceToken = NatPunchRequestTokenCodec.Encode(
NatPunchPeerRole.HostPresence,
session.HostPresenceHandle,
session.HostPresenceCapability);
gameplayNetManager.NatPunchModule.SendNatIntroduceRequest(mediator, presenceToken);
```
For direct connections, let the SDK drive those tokens from the same caller-owned
LiteNetLib socket that carries gameplay. Ask the routing listener to create the
bound manager, then configure and start that caller-owned manager yourself. The
factory does not open a socket, and synchronized events must remain enabled:
```csharp
RendezvousNetListener networkEvents = new();
NetManager gameplayNetManager = networkEvents.CreateManager();
if (!gameplayNetManager.Start(0))
{
throw new InvalidOperationException("The gameplay UDP socket could not start.");
}
```
The host polls join invitations asynchronously; that method only queues a
snapshot and never calls the manager. `Poll()` is the sole SDK path that invokes
LiteNetLib and dispatches its synchronized callbacks. Call it once per game
frame on the thread that owns the manager:
```csharp
RendezvousJoinClient joins = new(http);
using RendezvousHostCoordinator host = new(
gameplayNetManager,
networkEvents,
mediatorEndPoint,
session,
joins);
// Run periodically from the game's normal async scheduling path.
await host.RefreshJoinAttemptsAsync(cancellationToken);
// Godot _Process, Update, or the equivalent main-thread frame callback.
host.Poll();
```
Do not also call `gameplayNetManager.PollEvents()` or
`gameplayNetManager.NatPunchModule.PollEvents()` when a coordinator owns polling.
The host coordinator refreshes host presence, punches for queued invitations,
validates the introduction ticket, and accepts the direct request. Subscribe to
`AttemptCompleted`; a `Connected` result is raised only after LiteNetLib reports
the accepted peer as connected. Register ordinary gameplay callbacks on
`networkEvents.GameplayEvents`; the routing listener reserves Rendezvous direct
requests for ticket validation and forwards every other callback normally.
The joining game first requests an attempt through the typed start API. It returns
exactly one issued attempt or one terminal service outcome, so service authority
is not confused with a later locally observed traversal failure:
```csharp
RendezvousConnectionStartResult start = await joins.CreateConnectionAttemptAsync(
createJoinRequest,
cancellationToken: cancellationToken);
if (start.Outcome is { } serviceOutcome)
{
ShowConnectionFailure(serviceOutcome.Kind, serviceOutcome.Category);
return;
}
CreateJoinAttemptResponse attempt = start.Attempt
?? throw new InvalidOperationException("The typed start result was invalid.");
using RendezvousClientCoordinator client = new(
gameplayNetManager,
networkEvents,
mediatorEndPoint,
attempt);
// Godot _Process, Update, or the equivalent main-thread frame callback.
client.Poll();
```
NAT introduction changes the client state to `Connecting`; it is not success.
Only a `Connected` outcome supplies `Peer`. Completion exposes a stable kind,
source, category, phase, and elapsed duration. The default HTTP silence, punch,
and direct-connect budgets are five, ten, and five seconds respectively; configure
them through `RendezvousClientOptions` and `RendezvousCoordinatorOptions` when a
game has measured reasons to do so. The signed attempt expiry is always the
absolute upper bound.
Call `Cancel()` and then `Poll()` for local cancellation, or
`CancelAsync(joins, cancellationToken)` to also revoke the service attempt.
Terminal client paths complete exactly once and release all event subscriptions,
so late packets and callbacks are inert. Disposing a coordinator never stops or
disposes the caller-owned manager and does not touch an in-flight peer; call
`Cancel()` followed by `Poll()` first when that peer must also be disconnected.
After terminal completion, reporting is explicit and safe to retry. It sends only
the authenticated outcome enum and a coarse elapsed bucket—never the endpoint,
exact duration, diagnostic text, metadata, or player identity:
```csharp
RendezvousClientResult<ReportConnectionOutcomeResponse> report =
await client.ReportOutcomeAsync(joins, cancellationToken);
```
An optional `DedicatedFallback` is copied from the authoritative listing into the
issued attempt and terminal outcome. A local deployment may replace it with
`RendezvousCoordinatorOptions.DedicatedFallbackOverride`. The SDK only returns
the endpoint; it never connects automatically. The game must explicitly decide
whether to use it and then connect and authenticate through its own gameplay
transport. If the outcome has no fallback, v1 offers no relay.
Lease renewal is explicit and caller-controlled: Lease renewal is explicit and caller-controlled:
@@ -58,4 +175,23 @@ it when hosting stops. Use `IRendezvousPublisherClient` and
`IRendezvousSessionBrowserClient` as injection seams in game tests. The SDK disposes `IRendezvousSessionBrowserClient` as injection seams in game tests. The SDK disposes
the requests and responses it creates but never disposes the supplied `HttpClient`. the requests and responses it creates but never disposes the supplied `HttpClient`.
See the repository's ADR 0007 for retry, paging, ownership, and failure semantics. The host-side `ConnectionTicketValidator` is a bounded, thread-safe one-time gate.
Authorize only tickets delivered by the authenticated Rendezvous introduction,
then consume the exact ticket presented by the direct LiteNetLib connection:
```csharp
using ConnectionTicketValidator tickets = new();
tickets.TryAuthorize(attemptId, expectedTicket, expiresAt);
ConnectionTicketConsumptionResult admission = tickets.Consume(
attemptId,
presentedTicket);
```
An `Accepted` ticket authorizes only this connection attempt. The game must still
apply its own player identity, capacity, ban, and gameplay admission rules. Revoke
the attempt on cancellation and dispose the validator during host shutdown so its
keyed ticket digests are zeroed.
See the repository's ADR 0007 for HTTP ownership/retry semantics, ADR 0008 for
join-capability and connection-ticket security semantics, and ADR 0010 for typed
outcomes, deadlines, reporting, and caller-owned fallback.
@@ -42,6 +42,10 @@ public static class RendezvousClientResult
public sealed class PublishedSession public sealed class PublishedSession
{ {
private readonly object _timingGate = new();
private DateTimeOffset _expiresAt;
private int _leaseRenewAfterSeconds;
internal PublishedSession(RegisterSessionResponse response) internal PublishedSession(RegisterSessionResponse response)
{ {
ListingId = response.ListingId; ListingId = response.ListingId;
@@ -49,8 +53,8 @@ public sealed class PublishedSession
LeaseToken = response.LeaseToken; LeaseToken = response.LeaseToken;
HostPresenceHandle = response.HostPresenceHandle; HostPresenceHandle = response.HostPresenceHandle;
HostPresenceCapability = response.HostPresenceCapability; HostPresenceCapability = response.HostPresenceCapability;
ExpiresAt = response.ExpiresAt; _expiresAt = response.ExpiresAt;
LeaseRenewAfterSeconds = response.LeaseRenewAfterSeconds; _leaseRenewAfterSeconds = response.LeaseRenewAfterSeconds;
HostPresenceRefreshAfterSeconds = response.HostPresenceRefreshAfterSeconds; HostPresenceRefreshAfterSeconds = response.HostPresenceRefreshAfterSeconds;
} }
@@ -59,8 +63,41 @@ public sealed class PublishedSession
public string LeaseToken { get; } public string LeaseToken { get; }
public MediationHandle HostPresenceHandle { get; } public MediationHandle HostPresenceHandle { get; }
public string HostPresenceCapability { get; } public string HostPresenceCapability { get; }
public DateTimeOffset ExpiresAt { get; internal set; } public DateTimeOffset ExpiresAt
public int LeaseRenewAfterSeconds { get; internal set; } {
get
{
lock (_timingGate)
{
return _expiresAt;
}
}
internal set
{
lock (_timingGate)
{
_expiresAt = value;
}
}
}
public int LeaseRenewAfterSeconds
{
get
{
lock (_timingGate)
{
return _leaseRenewAfterSeconds;
}
}
internal set
{
lock (_timingGate)
{
_leaseRenewAfterSeconds = value;
}
}
}
public int HostPresenceRefreshAfterSeconds { get; } public int HostPresenceRefreshAfterSeconds { get; }
public override string ToString() => $"[PublishedSession {ListingId}; credentials redacted]"; public override string ToString() => $"[PublishedSession {ListingId}; credentials redacted]";
@@ -109,6 +146,38 @@ public interface IRendezvousSessionBrowserClient
CancellationToken cancellationToken = default); CancellationToken cancellationToken = default);
} }
public interface IRendezvousJoinClient
{
Task<RendezvousConnectionStartResult> CreateConnectionAttemptAsync(
CreateJoinAttemptRequest request,
NetworkEndpoint? dedicatedFallback = null,
CancellationToken cancellationToken = default);
Task<RendezvousClientResult<CreateJoinAttemptResponse>> CreateAsync(
CreateJoinAttemptRequest request,
CancellationToken cancellationToken = default);
Task<RendezvousClientResult<bool>> CancelAsync(
CreateJoinAttemptResponse attempt,
CancellationToken cancellationToken = default);
Task<RendezvousClientResult<BrowseHostJoinAttemptsResponse>> BrowseForHostAsync(
PublishedSession session,
int pageSize = ContractLimits.BrowserPageMaxItems,
string? cursor = null,
CancellationToken cancellationToken = default);
Task<RendezvousClientResult<IReadOnlyList<HostJoinAttempt>>> BrowseAllForHostAsync(
PublishedSession session,
int maximumPages = 100,
CancellationToken cancellationToken = default);
Task<RendezvousClientResult<ReportConnectionOutcomeResponse>> ReportOutcomeAsync(
CreateJoinAttemptResponse attempt,
RendezvousConnectionOutcome outcome,
CancellationToken cancellationToken = default);
}
public interface IRendezvousDelay public interface IRendezvousDelay
{ {
Task DelayAsync(TimeSpan delay, CancellationToken cancellationToken); Task DelayAsync(TimeSpan delay, CancellationToken cancellationToken);
@@ -117,6 +186,7 @@ public interface IRendezvousDelay
public sealed class RendezvousClientOptions public sealed class RendezvousClientOptions
{ {
public int MaximumSafeRetries { get; set; } = 2; public int MaximumSafeRetries { get; set; } = 2;
public TimeSpan RequestTimeout { get; set; } = TimeSpan.FromSeconds(5);
public TimeSpan InitialRetryDelay { get; set; } = TimeSpan.FromMilliseconds(200); public TimeSpan InitialRetryDelay { get; set; } = TimeSpan.FromMilliseconds(200);
public TimeSpan MaximumRetryDelay { get; set; } = TimeSpan.FromSeconds(2); public TimeSpan MaximumRetryDelay { get; set; } = TimeSpan.FromSeconds(2);
public double JitterRatio { get; set; } = 0.2; public double JitterRatio { get; set; } = 0.2;
@@ -124,6 +194,8 @@ public sealed class RendezvousClientOptions
internal void Validate() internal void Validate()
{ {
if (MaximumSafeRetries is < 0 or > 5 if (MaximumSafeRetries is < 0 or > 5
|| RequestTimeout <= TimeSpan.Zero
|| RequestTimeout > TimeSpan.FromSeconds(30)
|| InitialRetryDelay < TimeSpan.Zero || InitialRetryDelay < TimeSpan.Zero
|| MaximumRetryDelay < InitialRetryDelay || MaximumRetryDelay < InitialRetryDelay
|| MaximumRetryDelay > TimeSpan.FromSeconds(30) || MaximumRetryDelay > TimeSpan.FromSeconds(30)
@@ -139,3 +211,15 @@ internal sealed class SystemRendezvousDelay : IRendezvousDelay
public Task DelayAsync(TimeSpan delay, CancellationToken cancellationToken) => public Task DelayAsync(TimeSpan delay, CancellationToken cancellationToken) =>
Task.Delay(delay, cancellationToken); Task.Delay(delay, cancellationToken);
} }
internal static class RendezvousEndpoint
{
internal static NetworkEndpoint? Copy(NetworkEndpoint? endpoint) => endpoint is null
? null
: new NetworkEndpoint
{
AddressFamily = endpoint.AddressFamily,
Address = endpoint.Address,
Port = endpoint.Port,
};
}
@@ -24,6 +24,7 @@ internal sealed class RendezvousHttpTransport
_options = new RendezvousClientOptions _options = new RendezvousClientOptions
{ {
MaximumSafeRetries = suppliedOptions.MaximumSafeRetries, MaximumSafeRetries = suppliedOptions.MaximumSafeRetries,
RequestTimeout = suppliedOptions.RequestTimeout,
InitialRetryDelay = suppliedOptions.InitialRetryDelay, InitialRetryDelay = suppliedOptions.InitialRetryDelay,
MaximumRetryDelay = suppliedOptions.MaximumRetryDelay, MaximumRetryDelay = suppliedOptions.MaximumRetryDelay,
JitterRatio = suppliedOptions.JitterRatio, JitterRatio = suppliedOptions.JitterRatio,
@@ -38,11 +39,15 @@ internal sealed class RendezvousHttpTransport
for (int attempt = 0; ; attempt++) for (int attempt = 0; ; attempt++)
{ {
cancellationToken.ThrowIfCancellationRequested(); cancellationToken.ThrowIfCancellationRequested();
using CancellationTokenSource requestTimeout =
CancellationTokenSource.CreateLinkedTokenSource(cancellationToken);
requestTimeout.CancelAfter(_options.RequestTimeout);
CancellationToken requestCancellation = requestTimeout.Token;
try try
{ {
using HttpRequestMessage request = requestFactory(); using HttpRequestMessage request = requestFactory();
using HttpResponseMessage response = await _httpClient using HttpResponseMessage response = await _httpClient
.SendAsync(request, HttpCompletionOption.ResponseHeadersRead, cancellationToken) .SendAsync(request, HttpCompletionOption.ResponseHeadersRead, requestCancellation)
.ConfigureAwait(false); .ConfigureAwait(false);
if (response.IsSuccessStatusCode) if (response.IsSuccessStatusCode)
{ {
@@ -54,7 +59,7 @@ internal sealed class RendezvousHttpTransport
byte[] payload; byte[] payload;
try try
{ {
payload = await ReadBoundedAsync(response.Content, cancellationToken) payload = await ReadBoundedAsync(response.Content, requestCancellation)
.ConfigureAwait(false); .ConfigureAwait(false);
} }
catch (InvalidDataException) catch (InvalidDataException)
@@ -81,7 +86,7 @@ internal sealed class RendezvousHttpTransport
: RendezvousClientResult.Success(value); : RendezvousClientResult.Success(value);
} }
ApiError error = await ReadErrorAsync(response, cancellationToken).ConfigureAwait(false); ApiError error = await ReadErrorAsync(response, requestCancellation).ConfigureAwait(false);
int? retryAfter = error.RetryAfterSeconds ?? GetRetryAfterSeconds(response.Headers.RetryAfter); int? retryAfter = error.RetryAfterSeconds ?? GetRetryAfterSeconds(response.Headers.RetryAfter);
if (attempt < _options.MaximumSafeRetries && IsTransient(error.Code)) if (attempt < _options.MaximumSafeRetries && IsTransient(error.Code))
{ {
@@ -88,6 +88,7 @@ public sealed class RendezvousPublisherClient : IRendezvousPublisherClient
DisplayName = request.DisplayName, DisplayName = request.DisplayName,
Capacity = CopyCapacity(request.Capacity), Capacity = CopyCapacity(request.Capacity),
Metadata = CopyMetadata(request.Metadata), Metadata = CopyMetadata(request.Metadata),
DedicatedFallback = RendezvousEndpoint.Copy(request.DedicatedFallback),
}; };
return _transport.SendSafeAsync<bool>( return _transport.SendSafeAsync<bool>(
() => RendezvousHttpTransport.JsonRequest( () => RendezvousHttpTransport.JsonRequest(
@@ -138,6 +139,7 @@ public sealed class RendezvousPublisherClient : IRendezvousPublisherClient
Visibility = request.Visibility, Visibility = request.Visibility,
Capacity = CopyCapacity(request.Capacity), Capacity = CopyCapacity(request.Capacity),
Metadata = CopyMetadata(request.Metadata), Metadata = CopyMetadata(request.Metadata),
DedicatedFallback = RendezvousEndpoint.Copy(request.DedicatedFallback),
}; };
private static SessionCapacity CopyCapacity(SessionCapacity capacity) => new() private static SessionCapacity CopyCapacity(SessionCapacity capacity) => new()
@@ -148,4 +150,5 @@ public sealed class RendezvousPublisherClient : IRendezvousPublisherClient
private static Dictionary<string, string> CopyMetadata(Dictionary<string, string> metadata) => private static Dictionary<string, string> CopyMetadata(Dictionary<string, string> metadata) =>
new(metadata, StringComparer.Ordinal); new(metadata, StringComparer.Ordinal);
} }
@@ -0,0 +1,83 @@
using System.Text;
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Client;
public sealed class DirectConnectionRequest
{
public JoinAttemptId AttemptId { get; set; }
public string ConnectionTicket { get; set; } = string.Empty;
public override string ToString() =>
$"[DirectConnectionRequest {AttemptId}; ticket redacted]";
}
public static class DirectConnectionRequestCodec
{
public const int EncodedLength = 63;
private const int MagicLength = 4;
private const int AttemptIdLength = 16;
private const int TicketLength = ContractLimits.DerivedCredentialCharacters;
private static readonly byte[] Magic = [(byte)'R', (byte)'V', (byte)'D', (byte)'1'];
public static bool IsRendezvousRequest(ReadOnlySpan<byte> encoded) =>
encoded.Length >= MagicLength && encoded[..MagicLength].SequenceEqual(Magic);
public static byte[] Encode(JoinAttemptId attemptId, string connectionTicket)
{
if (attemptId.Value == Guid.Empty
|| connectionTicket is null
|| connectionTicket.Length != TicketLength
|| !ContractValidation.IsConnectionTicketValid(connectionTicket))
{
throw new ArgumentException("The direct connection request fields are invalid.");
}
byte[] encoded = new byte[EncodedLength];
Magic.CopyTo(encoded, 0);
if (!attemptId.Value.TryWriteBytes(encoded.AsSpan(MagicLength, AttemptIdLength)))
{
throw new InvalidOperationException("The join attempt identifier could not be encoded.");
}
Encoding.ASCII.GetBytes(
connectionTicket,
0,
connectionTicket.Length,
encoded,
MagicLength + AttemptIdLength);
return encoded;
}
public static bool TryDecode(
ReadOnlySpan<byte> encoded,
out DirectConnectionRequest? request)
{
request = null;
if (encoded.Length != EncodedLength
|| !encoded[..MagicLength].SequenceEqual(Magic))
{
return false;
}
Guid attemptId = new(encoded.Slice(MagicLength, AttemptIdLength));
if (attemptId == Guid.Empty)
{
return false;
}
string ticket = Encoding.ASCII.GetString(encoded[(MagicLength + AttemptIdLength)..]);
if (!ContractValidation.IsConnectionTicketValid(ticket))
{
return false;
}
request = new DirectConnectionRequest
{
AttemptId = new JoinAttemptId(attemptId),
ConnectionTicket = ticket,
};
return true;
}
}
@@ -0,0 +1,462 @@
using System.Net;
using System.Net.Sockets;
using FinalFactory.Rendezvous.Contracts;
using LiteNetLib;
namespace FinalFactory.Rendezvous.Client;
public sealed class RendezvousClientCoordinator : IDisposable
{
private readonly NetManager _manager;
private readonly RendezvousNetListener _networkEvents;
private readonly EventBasedNatPunchListener _punchEvents;
private readonly IPEndPoint _mediator;
private readonly CreateJoinAttemptResponse _attempt;
private readonly IRendezvousCoordinatorClock _clock;
private readonly RendezvousCoordinatorOptions _options;
private readonly RendezvousPunchRetrySchedule _retry;
private readonly object _completionGate = new();
private readonly TimeSpan _startedAt;
private readonly TimeSpan _attemptDeadline;
private readonly TimeSpan _punchDeadline;
private readonly NetworkEndpoint? _dedicatedFallback;
private NetPeer? _connectingPeer;
private IPEndPoint? _directEndpoint;
private TimeSpan? _directDeadline;
private RendezvousConnectionOutcome? _outcome;
private bool _cancelRequested;
private int _polling;
private bool _subscriptionsReleased;
private int _disposed;
public RendezvousClientCoordinator(
NetManager manager,
RendezvousNetListener networkEvents,
IPEndPoint mediator,
CreateJoinAttemptResponse attempt,
RendezvousCoordinatorOptions? options = null)
: this(
manager,
networkEvents,
mediator,
attempt,
options,
new SystemRendezvousCoordinatorClock())
{
}
internal RendezvousClientCoordinator(
NetManager manager,
RendezvousNetListener networkEvents,
IPEndPoint mediator,
CreateJoinAttemptResponse attempt,
RendezvousCoordinatorOptions? options,
IRendezvousCoordinatorClock clock)
{
_manager = manager ?? throw new ArgumentNullException(nameof(manager));
_networkEvents = networkEvents ?? throw new ArgumentNullException(nameof(networkEvents));
_punchEvents = _networkEvents.PunchEvents;
_mediator = mediator ?? throw new ArgumentNullException(nameof(mediator));
_attempt = attempt ?? throw new ArgumentNullException(nameof(attempt));
_clock = clock ?? throw new ArgumentNullException(nameof(clock));
_options = (options ?? new RendezvousCoordinatorOptions())
.CopyAndValidate();
_retry = new(_options, _clock);
RendezvousManagerGuard.Validate(_manager, _networkEvents);
DateTimeOffset startedUtc = _clock.UtcNow;
if (_mediator.Port is < 1 or > 65_535
|| _attempt.AttemptId.Value == Guid.Empty
|| _attempt.MediationHandle.Value == Guid.Empty
|| !ContractValidation.IsCapabilityValid(_attempt.ClientPunchCapability)
|| !ContractValidation.IsConnectionTicketValid(_attempt.ConnectionTicketDigest)
|| _attempt.ExpiresAt <= startedUtc)
{
throw new ArgumentException("The client traversal inputs are invalid.");
}
_startedAt = _clock.Elapsed;
_attemptDeadline = _startedAt + (_attempt.ExpiresAt - startedUtc);
_punchDeadline = Min(_attemptDeadline, _startedAt + _options.PunchTimeout);
_dedicatedFallback = RendezvousEndpoint.Copy(
_options.DedicatedFallbackOverride ?? _attempt.DedicatedFallback);
_networkEvents.RendezvousPeerConnected += OnPeerConnected;
_networkEvents.RendezvousPeerDisconnected += OnPeerDisconnected;
_networkEvents.RendezvousNetworkError += OnNetworkError;
_punchEvents.NatIntroductionSuccess += OnNatIntroductionSuccess;
}
public event EventHandler<RendezvousConnectionCompletedEventArgs>? Completed;
public RendezvousConnectionState State { get; private set; } = RendezvousConnectionState.Punching;
public NetPeer? ConnectedPeer { get; private set; }
public RendezvousConnectionOutcome? Outcome => Volatile.Read(ref _outcome);
public bool IsCompleted => Outcome is not null;
public void Cancel() => Volatile.Write(ref _cancelRequested, true);
public async Task<RendezvousClientResult<bool>> CancelAsync(
IRendezvousJoinClient joinClient,
CancellationToken cancellationToken = default)
{
if (joinClient is null)
{
throw new ArgumentNullException(nameof(joinClient));
}
ThrowIfDisposed();
Cancel();
return await joinClient.CancelAsync(_attempt, cancellationToken).ConfigureAwait(false);
}
public Task<RendezvousClientResult<ReportConnectionOutcomeResponse>> ReportOutcomeAsync(
IRendezvousJoinClient joinClient,
CancellationToken cancellationToken = default)
{
if (joinClient is null)
{
throw new ArgumentNullException(nameof(joinClient));
}
ThrowIfDisposed();
if (Outcome is null)
{
throw new InvalidOperationException("The connection attempt has not completed.");
}
return joinClient.ReportOutcomeAsync(_attempt, Outcome, cancellationToken);
}
public void Poll()
{
ThrowIfDisposed();
if (IsCompleted)
{
return;
}
if (Interlocked.Exchange(ref _polling, 1) != 0)
{
throw new InvalidOperationException("The Rendezvous coordinator cannot be polled concurrently or recursively.");
}
try
{
if (Volatile.Read(ref _cancelRequested))
{
DisconnectPendingPeer();
Complete(
RendezvousConnectionState.Cancelled,
ConnectionOutcomeKind.Cancelled,
RendezvousConnectionOutcomeSource.Caller,
RendezvousConnectionFailureCategory.Lifecycle,
CurrentPhase());
return;
}
if (!_manager.IsRunning)
{
CompleteManagerStopped();
return;
}
_manager.PollEvents();
_manager.NatPunchModule.PollEvents();
if (IsCompleted)
{
return;
}
DateTimeOffset now = _clock.UtcNow;
TimeSpan elapsed = _clock.Elapsed;
if (Volatile.Read(ref _cancelRequested))
{
DisconnectPendingPeer();
Complete(
RendezvousConnectionState.Cancelled,
ConnectionOutcomeKind.Cancelled,
RendezvousConnectionOutcomeSource.Caller,
RendezvousConnectionFailureCategory.Lifecycle,
CurrentPhase());
}
else if (!_manager.IsRunning)
{
CompleteManagerStopped();
}
else if (now >= _attempt.ExpiresAt || elapsed >= _attemptDeadline)
{
DisconnectPendingPeer();
Complete(
RendezvousConnectionState.TimedOut,
ConnectionOutcomeKind.AttemptExpired,
RendezvousConnectionOutcomeSource.RendezvousService,
RendezvousConnectionFailureCategory.Authorization,
RendezvousConnectionPhase.Authorization);
}
else if (State == RendezvousConnectionState.Punching)
{
if (elapsed >= _punchDeadline
|| _retry.IsExhausted && _retry.IsDue(elapsed))
{
Complete(
RendezvousConnectionState.TimedOut,
ConnectionOutcomeKind.PunchTimedOut,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.NatTraversal,
RendezvousConnectionPhase.NatTraversal);
return;
}
if (_retry.IsDue(elapsed))
{
_manager.NatPunchModule.SendNatIntroduceRequest(
_mediator,
NatPunchRequestTokenCodec.Encode(
NatPunchPeerRole.Client,
_attempt.MediationHandle,
_attempt.ClientPunchCapability));
_retry.RecordRequest();
}
}
else if (State == RendezvousConnectionState.Connecting
&& _directDeadline is TimeSpan directDeadline
&& directDeadline <= elapsed)
{
DisconnectPendingPeer();
Complete(
RendezvousConnectionState.TimedOut,
ConnectionOutcomeKind.DirectConnectTimedOut,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection);
}
}
finally
{
Volatile.Write(ref _polling, 0);
}
}
public void Dispose()
{
if (Interlocked.Exchange(ref _disposed, 1) != 0)
{
return;
}
if (!IsCompleted)
{
Complete(
RendezvousConnectionState.Disposed,
ConnectionOutcomeKind.Disposed,
RendezvousConnectionOutcomeSource.Lifecycle,
RendezvousConnectionFailureCategory.Lifecycle,
CurrentPhase());
}
ReleaseSubscriptions();
}
public override string ToString() =>
$"[RendezvousClientCoordinator {_attempt.AttemptId}; credentials redacted]";
private void OnNatIntroductionSuccess(
IPEndPoint target,
NatAddressType addressType,
string encodedIntroduction)
{
_ = addressType;
if (State != RendezvousConnectionState.Punching
|| !NatIntroductionTokenCodec.TryDecode(
encodedIntroduction,
out NatIntroductionToken? introduction)
|| introduction is null
|| introduction.AttemptId != _attempt.AttemptId
|| !NatIntroductionTokenCodec.MatchesDigest(
introduction.ConnectionTicket,
_attempt.ConnectionTicketDigest))
{
return;
}
byte[] connectionData = DirectConnectionRequestCodec.Encode(
introduction.AttemptId,
introduction.ConnectionTicket);
_directEndpoint = target;
_connectingPeer = _manager.Connect(target, connectionData);
if (_connectingPeer is null
|| _connectingPeer.ConnectionState != ConnectionState.Outgoing)
{
_connectingPeer = null;
Complete(
RendezvousConnectionState.Rejected,
ConnectionOutcomeKind.TransportError,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection);
return;
}
State = RendezvousConnectionState.Connecting;
_directDeadline = Min(
_attemptDeadline,
_clock.Elapsed + _options.DirectConnectTimeout);
}
private void OnPeerConnected(NetPeer peer)
{
if (State != RendezvousConnectionState.Connecting
|| !ReferenceEquals(peer, _connectingPeer))
{
return;
}
Complete(
RendezvousConnectionState.Connected,
ConnectionOutcomeKind.Connected,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.None,
RendezvousConnectionPhase.Complete,
peer);
}
private void OnPeerDisconnected(NetPeer peer, DisconnectInfo disconnectInfo)
{
if (State == RendezvousConnectionState.Connecting
&& ReferenceEquals(peer, _connectingPeer))
{
ConnectionOutcomeKind kind = disconnectInfo.Reason == DisconnectReason.Timeout
? ConnectionOutcomeKind.DirectConnectTimedOut
: disconnectInfo.Reason == DisconnectReason.ConnectionFailed
? ConnectionOutcomeKind.TransportError
: ConnectionOutcomeKind.HostRejected;
Complete(
kind == ConnectionOutcomeKind.DirectConnectTimedOut
? RendezvousConnectionState.TimedOut
: RendezvousConnectionState.Rejected,
kind,
kind == ConnectionOutcomeKind.HostRejected
? RendezvousConnectionOutcomeSource.RemoteHost
: RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection);
}
}
private void OnNetworkError(IPEndPoint endpoint, SocketError socketError)
{
_ = socketError;
if (State == RendezvousConnectionState.Punching && endpoint.Equals(_mediator))
{
Complete(
RendezvousConnectionState.Rejected,
ConnectionOutcomeKind.MediatorUnavailable,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.Mediation,
RendezvousConnectionPhase.Mediation);
}
else if (State == RendezvousConnectionState.Connecting
&& endpoint.Equals(_directEndpoint))
{
DisconnectPendingPeer();
Complete(
RendezvousConnectionState.Rejected,
ConnectionOutcomeKind.TransportError,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection);
}
}
private void DisconnectPendingPeer()
{
if (_connectingPeer is not null && State == RendezvousConnectionState.Connecting)
{
_connectingPeer.Disconnect();
}
}
private void Complete(
RendezvousConnectionState terminalState,
ConnectionOutcomeKind kind,
RendezvousConnectionOutcomeSource source,
RendezvousConnectionFailureCategory category,
RendezvousConnectionPhase phase,
NetPeer? peer = null)
{
RendezvousConnectionCompletedEventArgs completion;
lock (_completionGate)
{
if (_outcome is not null)
{
return;
}
RendezvousConnectionOutcome outcome = RendezvousConnectionOutcome.Create(
kind,
source,
category,
phase,
_clock.Elapsed - _startedAt,
ShouldOfferFallback(kind) ? _dedicatedFallback : null,
peer);
State = terminalState;
if (kind == ConnectionOutcomeKind.Connected)
{
ConnectedPeer = peer;
}
Volatile.Write(ref _outcome, outcome);
ReleaseSubscriptions();
completion = new(terminalState, outcome);
}
Completed?.Invoke(this, completion);
}
private void ReleaseSubscriptions()
{
lock (_completionGate)
{
if (_subscriptionsReleased)
{
return;
}
_networkEvents.RendezvousPeerConnected -= OnPeerConnected;
_networkEvents.RendezvousPeerDisconnected -= OnPeerDisconnected;
_networkEvents.RendezvousNetworkError -= OnNetworkError;
_punchEvents.NatIntroductionSuccess -= OnNatIntroductionSuccess;
_subscriptionsReleased = true;
}
}
private void CompleteManagerStopped() => Complete(
RendezvousConnectionState.ManagerStopped,
ConnectionOutcomeKind.ManagerStopped,
RendezvousConnectionOutcomeSource.Lifecycle,
RendezvousConnectionFailureCategory.Lifecycle,
CurrentPhase());
private RendezvousConnectionPhase CurrentPhase() => State switch
{
RendezvousConnectionState.Punching => RendezvousConnectionPhase.NatTraversal,
RendezvousConnectionState.Connecting => RendezvousConnectionPhase.DirectConnection,
_ => RendezvousConnectionPhase.Complete,
};
private static bool ShouldOfferFallback(ConnectionOutcomeKind kind) => kind is not (
ConnectionOutcomeKind.Connected
or ConnectionOutcomeKind.Cancelled
or ConnectionOutcomeKind.Disposed);
private static TimeSpan Min(TimeSpan left, TimeSpan right) =>
left <= right ? left : right;
private void ThrowIfDisposed()
{
if (Volatile.Read(ref _disposed) != 0)
{
throw new ObjectDisposedException(nameof(RendezvousClientCoordinator));
}
}
}
@@ -0,0 +1,228 @@
using System.Diagnostics;
using System.Security.Cryptography;
using FinalFactory.Rendezvous.Contracts;
using LiteNetLib;
namespace FinalFactory.Rendezvous.Client;
public enum RendezvousConnectionState
{
Punching = 1,
Connecting = 2,
Connected = 3,
Cancelled = 4,
TimedOut = 5,
Rejected = 6,
ManagerStopped = 7,
Disposed = 8,
}
public sealed class RendezvousConnectionCompletedEventArgs : EventArgs
{
[Obsolete("Completion events now expose a typed Outcome. Construct these arguments only for legacy test doubles.")]
public RendezvousConnectionCompletedEventArgs(
RendezvousConnectionState state,
NetPeer? peer)
: this(state, RendezvousCompletionInvariant.FromLegacy(state, peer))
{
}
internal RendezvousConnectionCompletedEventArgs(
RendezvousConnectionState state,
RendezvousConnectionOutcome outcome)
{
RendezvousCompletionInvariant.Validate(state, outcome);
State = state;
Outcome = outcome;
}
public RendezvousConnectionState State { get; }
public RendezvousConnectionOutcome Outcome { get; }
public NetPeer? Peer => Outcome.Peer;
}
internal static class RendezvousCompletionInvariant
{
internal static RendezvousConnectionOutcome FromLegacy(
RendezvousConnectionState state,
NetPeer? peer) => state switch
{
RendezvousConnectionState.Connected when peer is not null => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.Connected,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.None,
RendezvousConnectionPhase.Complete,
TimeSpan.Zero,
peer: peer),
RendezvousConnectionState.Cancelled => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.Cancelled,
RendezvousConnectionOutcomeSource.Caller,
RendezvousConnectionFailureCategory.Lifecycle,
RendezvousConnectionPhase.Complete,
TimeSpan.Zero),
RendezvousConnectionState.TimedOut => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.DirectConnectTimedOut,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection,
TimeSpan.Zero),
RendezvousConnectionState.Rejected => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.HostRejected,
RendezvousConnectionOutcomeSource.RemoteHost,
RendezvousConnectionFailureCategory.Authorization,
RendezvousConnectionPhase.Authorization,
TimeSpan.Zero),
RendezvousConnectionState.ManagerStopped => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.ManagerStopped,
RendezvousConnectionOutcomeSource.Lifecycle,
RendezvousConnectionFailureCategory.Lifecycle,
RendezvousConnectionPhase.Complete,
TimeSpan.Zero),
RendezvousConnectionState.Disposed => RendezvousConnectionOutcome.Create(
ConnectionOutcomeKind.Disposed,
RendezvousConnectionOutcomeSource.Lifecycle,
RendezvousConnectionFailureCategory.Lifecycle,
RendezvousConnectionPhase.Complete,
TimeSpan.Zero),
RendezvousConnectionState.Connected => throw new ArgumentNullException(
nameof(peer),
"A connected completion requires a peer."),
_ => throw new ArgumentOutOfRangeException(
nameof(state),
state,
"A completion event requires a terminal connection state."),
};
internal static void Validate(
RendezvousConnectionState state,
RendezvousConnectionOutcome outcome)
{
if (outcome is null)
{
throw new ArgumentNullException(nameof(outcome));
}
if ((state == RendezvousConnectionState.Connected) != outcome.IsSuccess)
{
throw new ArgumentException(
"The connection state and typed outcome contradict each other.",
nameof(outcome));
}
}
}
public sealed class RendezvousCoordinatorOptions
{
public int MaximumPunchRequests { get; set; } = 5;
public int MaximumAttemptChecksPerPoll { get; set; } = 128;
public TimeSpan InitialPunchRetryDelay { get; set; } = TimeSpan.FromMilliseconds(200);
public TimeSpan MaximumPunchRetryDelay { get; set; } = TimeSpan.FromSeconds(2);
public TimeSpan PunchTimeout { get; set; } = TimeSpan.FromSeconds(10);
public TimeSpan DirectConnectTimeout { get; set; } = TimeSpan.FromSeconds(5);
public TimeSpan ConnectionTicketLifetime { get; set; } = TimeSpan.FromSeconds(20);
public double JitterRatio { get; set; } = 0.2;
public NetworkEndpoint? DedicatedFallbackOverride { get; set; }
internal RendezvousCoordinatorOptions CopyAndValidate()
{
if (MaximumPunchRequests is < 1 or > 20
|| MaximumAttemptChecksPerPoll is < 1 or > 1_024
|| InitialPunchRetryDelay < TimeSpan.FromMilliseconds(10)
|| MaximumPunchRetryDelay < InitialPunchRetryDelay
|| MaximumPunchRetryDelay > TimeSpan.FromSeconds(10)
|| PunchTimeout <= TimeSpan.Zero
|| PunchTimeout > TimeSpan.FromSeconds(30)
|| DirectConnectTimeout <= TimeSpan.Zero
|| DirectConnectTimeout > TimeSpan.FromSeconds(30)
|| ConnectionTicketLifetime <= TimeSpan.Zero
|| ConnectionTicketLifetime > TimeSpan.FromSeconds(20)
|| JitterRatio is < 0 or > 1
|| DedicatedFallbackOverride is not null
&& !ContractValidation.IsNetworkEndpointValid(DedicatedFallbackOverride))
{
throw new ArgumentOutOfRangeException(nameof(RendezvousCoordinatorOptions));
}
return new RendezvousCoordinatorOptions
{
MaximumPunchRequests = MaximumPunchRequests,
MaximumAttemptChecksPerPoll = MaximumAttemptChecksPerPoll,
InitialPunchRetryDelay = InitialPunchRetryDelay,
MaximumPunchRetryDelay = MaximumPunchRetryDelay,
PunchTimeout = PunchTimeout,
DirectConnectTimeout = DirectConnectTimeout,
ConnectionTicketLifetime = ConnectionTicketLifetime,
JitterRatio = JitterRatio,
DedicatedFallbackOverride = RendezvousEndpoint.Copy(DedicatedFallbackOverride),
};
}
}
internal interface IRendezvousCoordinatorClock
{
DateTimeOffset UtcNow { get; }
TimeSpan Elapsed { get; }
}
internal sealed class SystemRendezvousCoordinatorClock : IRendezvousCoordinatorClock
{
private readonly long _origin = Stopwatch.GetTimestamp();
public DateTimeOffset UtcNow => DateTimeOffset.UtcNow;
public TimeSpan Elapsed => TimeSpan.FromSeconds(
(Stopwatch.GetTimestamp() - _origin) / (double)Stopwatch.Frequency);
}
internal static class RendezvousManagerGuard
{
internal static void Validate(
NetManager manager,
RendezvousNetListener networkEvents)
{
networkEvents.ValidateManager(manager);
if (!manager.IsRunning)
{
throw new InvalidOperationException("The caller-owned LiteNetLib manager must be running.");
}
if (!manager.NatPunchEnabled
|| manager.UnsyncedEvents
|| manager.NatPunchModule.UnsyncedEvents)
{
throw new InvalidOperationException(
"The caller-owned manager must enable NAT punching and synchronized event dispatch.");
}
}
}
internal sealed class RendezvousPunchRetrySchedule(
RendezvousCoordinatorOptions options,
IRendezvousCoordinatorClock clock)
{
public int RequestsSent { get; private set; }
public TimeSpan NextRequestAt { get; private set; } = TimeSpan.Zero;
public bool IsExhausted => RequestsSent >= options.MaximumPunchRequests;
public bool IsDue(TimeSpan elapsed) => elapsed >= NextRequestAt;
public void RecordRequest()
{
int exponent = Math.Min(RequestsSent, 30);
RequestsSent++;
double milliseconds = Math.Min(
options.InitialPunchRetryDelay.TotalMilliseconds * Math.Pow(2, exponent),
options.MaximumPunchRetryDelay.TotalMilliseconds);
if (options.JitterRatio > 0)
{
Span<byte> random = stackalloc byte[1];
RandomNumberGenerator.Fill(random);
double unit = random[0] / 255d;
double multiplier = 1 - options.JitterRatio + (2 * options.JitterRatio * unit);
milliseconds = Math.Min(
milliseconds * multiplier,
options.MaximumPunchRetryDelay.TotalMilliseconds);
}
NextRequestAt = clock.Elapsed + TimeSpan.FromMilliseconds(milliseconds);
}
}
@@ -0,0 +1,813 @@
using System.Net;
using System.Net.Sockets;
using FinalFactory.Rendezvous.Contracts;
using LiteNetLib;
namespace FinalFactory.Rendezvous.Client;
public enum RendezvousHostState
{
Active = 1,
ManagerStopped = 2,
Disposed = 3,
}
public sealed class RendezvousHostAttemptCompletedEventArgs : EventArgs
{
[Obsolete("Completion events now expose a typed Outcome. Construct these arguments only for legacy test doubles.")]
public RendezvousHostAttemptCompletedEventArgs(
JoinAttemptId attemptId,
RendezvousConnectionState state,
NetPeer? peer)
: this(attemptId, state, RendezvousCompletionInvariant.FromLegacy(state, peer))
{
}
internal RendezvousHostAttemptCompletedEventArgs(
JoinAttemptId attemptId,
RendezvousConnectionState state,
RendezvousConnectionOutcome outcome)
{
if (attemptId.Value == Guid.Empty)
{
throw new ArgumentException("The completed attempt ID is invalid.", nameof(attemptId));
}
RendezvousCompletionInvariant.Validate(state, outcome);
AttemptId = attemptId;
State = state;
Outcome = outcome;
}
public JoinAttemptId AttemptId { get; }
public RendezvousConnectionState State { get; }
public RendezvousConnectionOutcome Outcome { get; }
public NetPeer? Peer => Outcome.Peer;
}
public sealed class RendezvousHostCoordinator : IDisposable
{
private readonly NetManager _manager;
private readonly RendezvousNetListener _networkEvents;
private readonly EventBasedNatPunchListener _punchEvents;
private readonly IPEndPoint _mediator;
private readonly PublishedSession _session;
private readonly IRendezvousJoinClient _joinClient;
private readonly RendezvousCoordinatorOptions _options;
private readonly IRendezvousCoordinatorClock _clock;
private readonly ConnectionTicketValidator _tickets;
private readonly Dictionary<JoinAttemptId, PendingHostAttempt> _attempts = [];
private readonly Dictionary<NetPeer, JoinAttemptId> _acceptedPeers = [];
private readonly Dictionary<JoinAttemptId, DeferredConnectionRequest> _deferredRequests = [];
private readonly Dictionary<JoinAttemptId, DateTimeOffset> _terminalAttempts = [];
private readonly Queue<JoinAttemptId> _attemptSchedule = [];
private readonly SortedDictionary<long, Queue<HostAttemptDeadline>> _deadlines = [];
private readonly List<JoinAttemptId> _cleanupScratch = [];
private HostJoinAttempt[]? _latestSnapshot;
private DateTimeOffset _nextPresenceAt = DateTimeOffset.MinValue;
private DateTimeOffset _nextTerminalCleanupAt = DateTimeOffset.MinValue;
private int _refreshing;
private int _polling;
private bool _subscriptionsReleased;
private int _disposed;
public RendezvousHostCoordinator(
NetManager manager,
RendezvousNetListener networkEvents,
IPEndPoint mediator,
PublishedSession session,
IRendezvousJoinClient joinClient,
RendezvousCoordinatorOptions? options = null)
: this(
manager,
networkEvents,
mediator,
session,
joinClient,
options,
new SystemRendezvousCoordinatorClock(),
null)
{
}
internal RendezvousHostCoordinator(
NetManager manager,
RendezvousNetListener networkEvents,
IPEndPoint mediator,
PublishedSession session,
IRendezvousJoinClient joinClient,
RendezvousCoordinatorOptions? options,
IRendezvousCoordinatorClock clock,
ConnectionTicketValidator? tickets)
{
_manager = manager ?? throw new ArgumentNullException(nameof(manager));
_networkEvents = networkEvents ?? throw new ArgumentNullException(nameof(networkEvents));
_punchEvents = _networkEvents.PunchEvents;
_mediator = mediator ?? throw new ArgumentNullException(nameof(mediator));
_session = session ?? throw new ArgumentNullException(nameof(session));
_joinClient = joinClient ?? throw new ArgumentNullException(nameof(joinClient));
_options = (options ?? new RendezvousCoordinatorOptions()).CopyAndValidate();
_clock = clock ?? throw new ArgumentNullException(nameof(clock));
_tickets = tickets ?? new ConnectionTicketValidator();
RendezvousManagerGuard.Validate(_manager, _networkEvents);
ValidateInputs();
_networkEvents.RendezvousConnectionRequest += OnConnectionRequest;
_networkEvents.RendezvousPeerConnected += OnPeerConnected;
_networkEvents.RendezvousPeerDisconnected += OnPeerDisconnected;
_networkEvents.RendezvousNetworkError += OnNetworkError;
_punchEvents.NatIntroductionSuccess += OnNatIntroductionSuccess;
}
public event EventHandler<RendezvousHostAttemptCompletedEventArgs>? AttemptCompleted;
public RendezvousHostState State { get; private set; } = RendezvousHostState.Active;
public int PendingAttemptCount => _attempts.Count;
internal int DeferredRequestCount => _deferredRequests.Count;
public async Task<RendezvousClientResult<int>> RefreshJoinAttemptsAsync(
CancellationToken cancellationToken = default)
{
ThrowIfDisposed();
if (Interlocked.Exchange(ref _refreshing, 1) != 0)
{
throw new InvalidOperationException("A host invitation refresh is already running.");
}
try
{
RendezvousClientResult<IReadOnlyList<HostJoinAttempt>> result =
await _joinClient.BrowseAllForHostAsync(
_session,
cancellationToken: cancellationToken).ConfigureAwait(false);
if (!result.IsSuccess || result.Value is null)
{
return RendezvousClientResult.Failure<int>(
result.Error,
result.Message,
result.RetryAfterSeconds);
}
HostJoinAttempt[] snapshot = result.Value.Select(CopyAttempt).ToArray();
if (Volatile.Read(ref _disposed) != 0)
{
throw new ObjectDisposedException(nameof(RendezvousHostCoordinator));
}
Interlocked.Exchange(ref _latestSnapshot, snapshot);
if (Volatile.Read(ref _disposed) != 0)
{
Interlocked.Exchange(ref _latestSnapshot, null);
throw new ObjectDisposedException(nameof(RendezvousHostCoordinator));
}
return RendezvousClientResult.Success(snapshot.Length);
}
finally
{
Volatile.Write(ref _refreshing, 0);
}
}
public void Poll()
{
ThrowIfDisposed();
if (State != RendezvousHostState.Active)
{
return;
}
if (Interlocked.Exchange(ref _polling, 1) != 0)
{
throw new InvalidOperationException("The Rendezvous coordinator cannot be polled concurrently or recursively.");
}
try
{
ApplySnapshots();
if (!_manager.IsRunning)
{
Stop(
RendezvousHostState.ManagerStopped,
RendezvousConnectionState.ManagerStopped,
ConnectionOutcomeKind.ManagerStopped);
return;
}
_manager.NatPunchModule.PollEvents();
_manager.PollEvents();
_manager.NatPunchModule.PollEvents();
if (State != RendezvousHostState.Active)
{
return;
}
DateTimeOffset now = _clock.UtcNow;
TimeSpan elapsed = _clock.Elapsed;
if (!_manager.IsRunning)
{
Stop(
RendezvousHostState.ManagerStopped,
RendezvousConnectionState.ManagerStopped,
ConnectionOutcomeKind.ManagerStopped);
return;
}
RefreshPresence(now);
ProcessDueDeadlines(elapsed);
if (State != RendezvousHostState.Active)
{
return;
}
int checks = Math.Min(
_attemptSchedule.Count,
_options.MaximumAttemptChecksPerPoll);
for (int index = 0; index < checks; index++)
{
JoinAttemptId attemptId = _attemptSchedule.Dequeue();
if (!_attempts.TryGetValue(attemptId, out PendingHostAttempt? attempt))
{
continue;
}
if (attempt.State != RendezvousConnectionState.Punching)
{
continue;
}
if (attempt.Retry.IsDue(elapsed))
{
if (attempt.Retry.IsExhausted)
{
CompleteAttempt(
attemptId,
RendezvousConnectionState.TimedOut,
ConnectionOutcomeKind.PunchTimedOut,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.NatTraversal,
RendezvousConnectionPhase.NatTraversal);
continue;
}
_manager.NatPunchModule.SendNatIntroduceRequest(
_mediator,
NatPunchRequestTokenCodec.Encode(
NatPunchPeerRole.Host,
attempt.Invitation.MediationHandle,
attempt.Invitation.HostPunchCapability));
attempt.Retry.RecordRequest();
}
_attemptSchedule.Enqueue(attemptId);
}
if (now >= _nextTerminalCleanupAt)
{
_cleanupScratch.Clear();
foreach (KeyValuePair<JoinAttemptId, DateTimeOffset> terminal in _terminalAttempts)
{
if (terminal.Value <= now)
{
_cleanupScratch.Add(terminal.Key);
}
}
foreach (JoinAttemptId attemptId in _cleanupScratch)
{
_terminalAttempts.Remove(attemptId);
}
_nextTerminalCleanupAt = now + TimeSpan.FromSeconds(1);
}
}
finally
{
Volatile.Write(ref _polling, 0);
}
}
public void Dispose()
{
if (Interlocked.Exchange(ref _disposed, 1) != 0)
{
return;
}
Stop(
RendezvousHostState.Disposed,
RendezvousConnectionState.Disposed,
ConnectionOutcomeKind.Disposed);
Interlocked.Exchange(ref _latestSnapshot, null);
_attemptSchedule.Clear();
_deadlines.Clear();
_terminalAttempts.Clear();
_cleanupScratch.Clear();
_tickets.Dispose();
}
public override string ToString() =>
$"[RendezvousHostCoordinator {_session.ListingId}; credentials redacted]";
private void ApplySnapshots()
{
HostJoinAttempt[]? latest = Interlocked.Exchange(ref _latestSnapshot, null);
if (latest is null)
{
return;
}
DateTimeOffset now = _clock.UtcNow;
TimeSpan elapsed = _clock.Elapsed;
foreach (HostJoinAttempt invitation in latest)
{
if (invitation.AttemptId.Value == Guid.Empty
|| invitation.MediationHandle.Value == Guid.Empty
|| !ContractValidation.IsCapabilityValid(invitation.HostPunchCapability)
|| !ContractValidation.IsConnectionTicketValid(
invitation.ConnectionTicketDigest))
{
continue;
}
if (invitation.IsCancelled)
{
if (_attempts.ContainsKey(invitation.AttemptId))
{
CompleteAttempt(
invitation.AttemptId,
RendezvousConnectionState.Cancelled,
ConnectionOutcomeKind.Cancelled,
RendezvousConnectionOutcomeSource.RendezvousService,
RendezvousConnectionFailureCategory.Lifecycle,
RendezvousConnectionPhase.Authorization);
}
_terminalAttempts[invitation.AttemptId] = invitation.ExpiresAt;
continue;
}
if (invitation.ExpiresAt <= now
|| _attempts.ContainsKey(invitation.AttemptId)
|| _terminalAttempts.ContainsKey(invitation.AttemptId))
{
continue;
}
TimeSpan attemptDeadline = elapsed + (invitation.ExpiresAt - now);
TimeSpan punchDeadline = Min(
attemptDeadline,
elapsed + _options.PunchTimeout);
_attempts.Add(
invitation.AttemptId,
new PendingHostAttempt(
CopyAttempt(invitation),
new RendezvousPunchRetrySchedule(_options, _clock),
elapsed,
attemptDeadline,
punchDeadline));
EnqueueDeadline(
new HostAttemptDeadline(
invitation.AttemptId,
RendezvousConnectionState.Punching,
punchDeadline));
_attemptSchedule.Enqueue(invitation.AttemptId);
}
}
private void RefreshPresence(DateTimeOffset now)
{
if (now < _nextPresenceAt || now >= _session.ExpiresAt)
{
return;
}
_manager.NatPunchModule.SendNatIntroduceRequest(
_mediator,
NatPunchRequestTokenCodec.Encode(
NatPunchPeerRole.HostPresence,
_session.HostPresenceHandle,
_session.HostPresenceCapability));
_nextPresenceAt = now + TimeSpan.FromSeconds(_session.HostPresenceRefreshAfterSeconds);
}
private void OnNatIntroductionSuccess(
IPEndPoint target,
NatAddressType addressType,
string encodedIntroduction)
{
_ = target;
_ = addressType;
if (!NatIntroductionTokenCodec.TryDecode(
encodedIntroduction,
out NatIntroductionToken? introduction)
|| introduction is null
|| !_attempts.TryGetValue(introduction.AttemptId, out PendingHostAttempt? attempt)
|| !NatIntroductionTokenCodec.MatchesDigest(
introduction.ConnectionTicket,
attempt.Invitation.ConnectionTicketDigest)
|| !_tickets.TryAuthorize(
introduction.AttemptId,
introduction.ConnectionTicket,
Min(
attempt.Invitation.ExpiresAt,
_clock.UtcNow + _options.ConnectionTicketLifetime)))
{
return;
}
attempt.State = RendezvousConnectionState.Connecting;
attempt.DirectDeadline = Min(
attempt.AttemptDeadline,
_clock.Elapsed + _options.DirectConnectTimeout);
EnqueueDeadline(new HostAttemptDeadline(
introduction.AttemptId,
RendezvousConnectionState.Connecting,
attempt.DirectDeadline.Value));
if (_deferredRequests.Remove(
introduction.AttemptId,
out DeferredConnectionRequest? deferred))
{
AcceptAuthorizedRequest(
introduction.AttemptId,
attempt,
deferred.Request,
deferred.ConnectionTicket);
}
}
private void OnConnectionRequest(ConnectionRequest request)
{
ReadOnlySpan<byte> data = request.Data.GetRemainingBytesSpan();
if (!DirectConnectionRequestCodec.IsRendezvousRequest(data))
{
return;
}
if (!DirectConnectionRequestCodec.TryDecode(data, out DirectConnectionRequest? connection)
|| connection is null
|| !_attempts.TryGetValue(connection.AttemptId, out PendingHostAttempt? attempt)
|| !NatIntroductionTokenCodec.MatchesDigest(
connection.ConnectionTicket,
attempt.Invitation.ConnectionTicketDigest))
{
request.RejectForce([]);
return;
}
if (attempt.State == RendezvousConnectionState.Punching)
{
_deferredRequests[connection.AttemptId] = new(
request,
connection.ConnectionTicket);
return;
}
if (attempt.State != RendezvousConnectionState.Connecting)
{
request.RejectForce([]);
return;
}
AcceptAuthorizedRequest(
connection.AttemptId,
attempt,
request,
connection.ConnectionTicket);
}
private void OnPeerConnected(NetPeer peer)
{
if (_acceptedPeers.TryGetValue(peer, out JoinAttemptId attemptId))
{
CompleteAttempt(
attemptId,
RendezvousConnectionState.Connected,
ConnectionOutcomeKind.Connected,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.None,
RendezvousConnectionPhase.Complete,
peer);
}
}
private void OnPeerDisconnected(NetPeer peer, DisconnectInfo disconnectInfo)
{
_ = disconnectInfo;
if (_acceptedPeers.TryGetValue(peer, out JoinAttemptId attemptId))
{
ConnectionOutcomeKind kind = disconnectInfo.Reason == DisconnectReason.Timeout
? ConnectionOutcomeKind.DirectConnectTimedOut
: ConnectionOutcomeKind.TransportError;
CompleteAttempt(
attemptId,
kind == ConnectionOutcomeKind.DirectConnectTimedOut
? RendezvousConnectionState.TimedOut
: RendezvousConnectionState.Rejected,
kind,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.DirectConnection,
RendezvousConnectionPhase.DirectConnection);
}
}
private void OnNetworkError(IPEndPoint endpoint, SocketError socketError)
{
_ = socketError;
if (!endpoint.Equals(_mediator))
{
return;
}
foreach (JoinAttemptId attemptId in _attempts
.Where(static item => item.Value.State == RendezvousConnectionState.Punching)
.Select(static item => item.Key)
.ToArray())
{
CompleteAttempt(
attemptId,
RendezvousConnectionState.Rejected,
ConnectionOutcomeKind.MediatorUnavailable,
RendezvousConnectionOutcomeSource.LocalTraversal,
RendezvousConnectionFailureCategory.Mediation,
RendezvousConnectionPhase.Mediation);
}
}
private void CompleteAttempt(
JoinAttemptId attemptId,
RendezvousConnectionState state,
ConnectionOutcomeKind kind,
RendezvousConnectionOutcomeSource source,
RendezvousConnectionFailureCategory category,
RendezvousConnectionPhase phase,
NetPeer? peer = null)
{
if (TryCompleteAttempt(
attemptId,
state,
kind,
source,
category,
phase,
peer,
out RendezvousHostAttemptCompletedEventArgs? completion))
{
AttemptCompleted?.Invoke(this, completion!);
}
}
private bool TryCompleteAttempt(
JoinAttemptId attemptId,
RendezvousConnectionState state,
ConnectionOutcomeKind kind,
RendezvousConnectionOutcomeSource source,
RendezvousConnectionFailureCategory category,
RendezvousConnectionPhase phase,
NetPeer? peer,
out RendezvousHostAttemptCompletedEventArgs? completion)
{
completion = null;
if (!_attempts.Remove(attemptId, out PendingHostAttempt? attempt))
{
return false;
}
if (attempt.AcceptedPeer is not null)
{
_acceptedPeers.Remove(attempt.AcceptedPeer);
if (kind != ConnectionOutcomeKind.Connected)
{
attempt.AcceptedPeer.Disconnect();
}
}
if (_deferredRequests.Remove(attemptId, out DeferredConnectionRequest? deferred))
{
deferred.Request.RejectForce([]);
}
_tickets.Revoke(attemptId);
_terminalAttempts[attemptId] = attempt.Invitation.ExpiresAt;
RendezvousConnectionOutcome outcome = RendezvousConnectionOutcome.Create(
kind,
source,
category,
phase,
_clock.Elapsed - attempt.StartedAt,
peer: peer);
completion = new(attemptId, state, outcome);
return true;
}
private void Stop(
RendezvousHostState hostState,
RendezvousConnectionState attemptState,
ConnectionOutcomeKind outcomeKind)
{
if (State != RendezvousHostState.Active)
{
return;
}
State = hostState;
List<RendezvousHostAttemptCompletedEventArgs> completions = [];
foreach (JoinAttemptId attemptId in _attempts.Keys.ToArray())
{
RendezvousConnectionPhase phase = _attempts[attemptId].State
== RendezvousConnectionState.Connecting
? RendezvousConnectionPhase.DirectConnection
: RendezvousConnectionPhase.NatTraversal;
if (TryCompleteAttempt(
attemptId,
attemptState,
outcomeKind,
RendezvousConnectionOutcomeSource.Lifecycle,
RendezvousConnectionFailureCategory.Lifecycle,
phase,
null,
out RendezvousHostAttemptCompletedEventArgs? completion))
{
completions.Add(completion!);
}
}
ReleaseSubscriptions();
foreach (RendezvousHostAttemptCompletedEventArgs completion in completions)
{
AttemptCompleted?.Invoke(this, completion);
}
}
private void ReleaseSubscriptions()
{
if (_subscriptionsReleased)
{
return;
}
_networkEvents.RendezvousConnectionRequest -= OnConnectionRequest;
_networkEvents.RendezvousPeerConnected -= OnPeerConnected;
_networkEvents.RendezvousPeerDisconnected -= OnPeerDisconnected;
_networkEvents.RendezvousNetworkError -= OnNetworkError;
_punchEvents.NatIntroductionSuccess -= OnNatIntroductionSuccess;
_subscriptionsReleased = true;
}
private void ValidateInputs()
{
if (_mediator.Port is < 1 or > 65_535
|| _session.HostPresenceHandle.Value == Guid.Empty
|| !ContractValidation.IsCapabilityValid(_session.HostPresenceCapability)
|| _session.HostPresenceRefreshAfterSeconds < 1
|| _session.ExpiresAt <= _clock.UtcNow)
{
throw new ArgumentException("The host traversal inputs are invalid.");
}
}
private static HostJoinAttempt CopyAttempt(HostJoinAttempt attempt) => new()
{
AttemptId = attempt.AttemptId,
MediationHandle = attempt.MediationHandle,
HostPunchCapability = attempt.HostPunchCapability,
ConnectionTicketDigest = attempt.ConnectionTicketDigest,
IsCancelled = attempt.IsCancelled,
ExpiresAt = attempt.ExpiresAt,
};
private static TimeSpan Min(TimeSpan left, TimeSpan right) =>
left <= right ? left : right;
private static DateTimeOffset Min(DateTimeOffset left, DateTimeOffset right) =>
left <= right ? left : right;
private void EnqueueDeadline(HostAttemptDeadline deadline)
{
if (!_deadlines.TryGetValue(deadline.Deadline.Ticks, out Queue<HostAttemptDeadline>? bucket))
{
bucket = new Queue<HostAttemptDeadline>();
_deadlines.Add(deadline.Deadline.Ticks, bucket);
}
bucket.Enqueue(deadline);
}
private void ProcessDueDeadlines(TimeSpan elapsed)
{
while (_deadlines.Count > 0)
{
KeyValuePair<long, Queue<HostAttemptDeadline>> first = _deadlines.First();
if (first.Key > elapsed.Ticks)
{
return;
}
HostAttemptDeadline deadline = first.Value.Dequeue();
if (first.Value.Count == 0)
{
_deadlines.Remove(first.Key);
}
if (!_attempts.TryGetValue(deadline.AttemptId, out PendingHostAttempt? attempt)
|| attempt.State != deadline.ExpectedState
|| (deadline.ExpectedState == RendezvousConnectionState.Punching
? attempt.PunchDeadline
: attempt.DirectDeadline) != deadline.Deadline)
{
continue;
}
bool expired = elapsed >= attempt.AttemptDeadline;
CompleteAttempt(
deadline.AttemptId,
RendezvousConnectionState.TimedOut,
expired
? ConnectionOutcomeKind.AttemptExpired
: deadline.ExpectedState == RendezvousConnectionState.Punching
? ConnectionOutcomeKind.PunchTimedOut
: ConnectionOutcomeKind.DirectConnectTimedOut,
expired
? RendezvousConnectionOutcomeSource.RendezvousService
: RendezvousConnectionOutcomeSource.LocalTraversal,
expired
? RendezvousConnectionFailureCategory.Authorization
: deadline.ExpectedState == RendezvousConnectionState.Punching
? RendezvousConnectionFailureCategory.NatTraversal
: RendezvousConnectionFailureCategory.DirectConnection,
expired
? RendezvousConnectionPhase.Authorization
: deadline.ExpectedState == RendezvousConnectionState.Punching
? RendezvousConnectionPhase.NatTraversal
: RendezvousConnectionPhase.DirectConnection);
if (State != RendezvousHostState.Active)
{
return;
}
}
}
private void AcceptAuthorizedRequest(
JoinAttemptId attemptId,
PendingHostAttempt attempt,
ConnectionRequest request,
string connectionTicket)
{
ConnectionTicketConsumptionResult consumption = _tickets.Consume(
attemptId,
connectionTicket);
if (consumption != ConnectionTicketConsumptionResult.Accepted)
{
request.RejectForce([]);
return;
}
NetPeer peer = request.Accept();
attempt.AcceptedPeer = peer;
_acceptedPeers[peer] = attemptId;
}
private void ThrowIfDisposed()
{
if (Volatile.Read(ref _disposed) != 0)
{
throw new ObjectDisposedException(nameof(RendezvousHostCoordinator));
}
}
private sealed class PendingHostAttempt(
HostJoinAttempt invitation,
RendezvousPunchRetrySchedule retry,
TimeSpan startedAt,
TimeSpan attemptDeadline,
TimeSpan punchDeadline)
{
internal HostJoinAttempt Invitation { get; } = invitation;
internal RendezvousPunchRetrySchedule Retry { get; } = retry;
internal TimeSpan StartedAt { get; } = startedAt;
internal TimeSpan AttemptDeadline { get; } = attemptDeadline;
internal TimeSpan PunchDeadline { get; } = punchDeadline;
internal TimeSpan? DirectDeadline { get; set; }
internal RendezvousConnectionState State { get; set; } = RendezvousConnectionState.Punching;
internal NetPeer? AcceptedPeer { get; set; }
}
private sealed class HostAttemptDeadline(
JoinAttemptId attemptId,
RendezvousConnectionState expectedState,
TimeSpan deadline)
{
internal JoinAttemptId AttemptId { get; } = attemptId;
internal RendezvousConnectionState ExpectedState { get; } = expectedState;
internal TimeSpan Deadline { get; } = deadline;
}
private sealed class DeferredConnectionRequest(
ConnectionRequest request,
string connectionTicket)
{
internal ConnectionRequest Request { get; } = request;
internal string ConnectionTicket { get; } = connectionTicket;
}
}
@@ -0,0 +1,114 @@
using System.Net;
using System.Net.Sockets;
using LiteNetLib;
using LiteNetLib.Utils;
namespace FinalFactory.Rendezvous.Client;
public sealed class RendezvousNetListener : INetEventListener
{
private NetManager? _manager;
public EventBasedNetListener GameplayEvents { get; } = new();
public EventBasedNatPunchListener PunchEvents { get; } = new();
public NetManager CreateManager()
{
if (_manager is not null)
{
throw new InvalidOperationException(
"This Rendezvous listener is already bound to a LiteNetLib manager.");
}
NetManager manager = new(this) { NatPunchEnabled = true };
manager.NatPunchModule.Init(PunchEvents);
_manager = manager;
return manager;
}
internal event Action<NetPeer>? RendezvousPeerConnected;
internal event Action<NetPeer, DisconnectInfo>? RendezvousPeerDisconnected;
internal event Action<ConnectionRequest>? RendezvousConnectionRequest;
internal event Action<IPEndPoint, SocketError>? RendezvousNetworkError;
internal void ValidateManager(NetManager manager)
{
if (!ReferenceEquals(_manager, manager))
{
throw new InvalidOperationException(
"The LiteNetLib manager must be created by this Rendezvous listener.");
}
}
public void OnPeerConnected(NetPeer peer)
{
RendezvousPeerConnected?.Invoke(peer);
((INetEventListener)GameplayEvents).OnPeerConnected(peer);
}
public void OnPeerDisconnected(NetPeer peer, DisconnectInfo disconnectInfo)
{
RendezvousPeerDisconnected?.Invoke(peer, disconnectInfo);
((INetEventListener)GameplayEvents).OnPeerDisconnected(peer, disconnectInfo);
}
public void OnNetworkError(IPEndPoint endPoint, SocketError socketError)
{
RendezvousNetworkError?.Invoke(endPoint, socketError);
((INetEventListener)GameplayEvents).OnNetworkError(endPoint, socketError);
}
public void OnNetworkReceive(
NetPeer peer,
NetPacketReader reader,
byte channelNumber,
DeliveryMethod deliveryMethod) =>
((INetEventListener)GameplayEvents).OnNetworkReceive(
peer,
reader,
channelNumber,
deliveryMethod);
public void OnNetworkReceiveUnconnected(
IPEndPoint remoteEndPoint,
NetPacketReader reader,
UnconnectedMessageType messageType) =>
((INetEventListener)GameplayEvents).OnNetworkReceiveUnconnected(
remoteEndPoint,
reader,
messageType);
public void OnNetworkLatencyUpdate(NetPeer peer, int latency) =>
((INetEventListener)GameplayEvents).OnNetworkLatencyUpdate(peer, latency);
public void OnConnectionRequest(ConnectionRequest request)
{
int position = request.Data.Position;
bool isRendezvous = DirectConnectionRequestCodec.IsRendezvousRequest(
request.Data.GetRemainingBytesSpan());
request.Data.SetPosition(position);
if (!isRendezvous)
{
((INetEventListener)GameplayEvents).OnConnectionRequest(request);
return;
}
Action<ConnectionRequest>? handler = RendezvousConnectionRequest;
if (handler is null)
{
request.RejectForce([]);
return;
}
handler(request);
}
public void OnMessageDelivered(NetPeer peer, object userData) =>
((INetEventListener)GameplayEvents).OnMessageDelivered(peer, userData);
public void OnNtpResponse(NtpPacket packet) =>
((INetEventListener)GameplayEvents).OnNtpResponse(packet);
public void OnPeerAddressChanged(NetPeer peer, IPEndPoint previousAddress) =>
((INetEventListener)GameplayEvents).OnPeerAddressChanged(peer, previousAddress);
}
@@ -4,7 +4,7 @@
".NETStandard,Version=v2.1": { ".NETStandard,Version=v2.1": {
"LiteNetLib": { "LiteNetLib": {
"type": "Direct", "type": "Direct",
"requested": "[2.1.4, )", "requested": "[2.1.4, 2.1.4]",
"resolved": "2.1.4", "resolved": "2.1.4",
"contentHash": "KWlxvMw3Urpqj9joD96LRiK+LC62pQNs/zkXRJc+rHnxgkGp+vV703xzDrxRmv+V1YhCFfIGzs5nrVWtREIlyA==" "contentHash": "KWlxvMw3Urpqj9joD96LRiK+LC62pQNs/zkXRJc+rHnxgkGp+vV703xzDrxRmv+V1YhCFfIGzs5nrVWtREIlyA=="
}, },
@@ -49,6 +49,27 @@ public enum ConnectionOutcomeKind
HostRejected = 7, HostRejected = 7,
TransportFailed = 8, TransportFailed = 8,
FallbackOffered = 9, FallbackOffered = 9,
DirectoryNotFound = 10,
AttemptExpired = 11,
Unauthorized = 12,
RateLimited = 13,
NoHostPresence = 14,
ServiceUnavailable = 15,
MediatorUnavailable = 16,
PunchTimedOut = 17,
DirectConnectTimedOut = 18,
TransportError = 19,
ManagerStopped = 20,
Disposed = 21,
}
public enum ConnectionElapsedBucket
{
UnderOneSecond = 1,
OneToFiveSeconds = 2,
FiveToFifteenSeconds = 3,
FifteenToThirtySeconds = 4,
ThirtySecondsOrMore = 5,
} }
public enum UdpPresenceMessageType : byte public enum UdpPresenceMessageType : byte
@@ -23,6 +23,8 @@ public static class ContractLimits
public const int OpaqueHttpCredentialMaxCharacters = 1_024; public const int OpaqueHttpCredentialMaxCharacters = 1_024;
public const int UdpCapabilityMaxCharacters = 192; public const int UdpCapabilityMaxCharacters = 192;
public const int ConnectionTicketMaxCharacters = 192; public const int ConnectionTicketMaxCharacters = 192;
public const int DerivedCredentialCharacters = 43;
public const int NatPunchRequestTokenCharacters = 192;
public const int LiteNetLibNatTokenMaxCharacters = 256; public const int LiteNetLibNatTokenMaxCharacters = 256;
public const int SessionCapacityMaxPlayers = 10_000; public const int SessionCapacityMaxPlayers = 10_000;
} }
@@ -45,6 +45,23 @@ public static class ContractValidation
public static bool IsDiagnosticCodeValid(string? value) => public static bool IsDiagnosticCodeValid(string? value) =>
value is null || IsVisibleAsciiWithin(value, ContractLimits.DiagnosticCodeMaxCharacters); value is null || IsVisibleAsciiWithin(value, ContractLimits.DiagnosticCodeMaxCharacters);
public static bool IsReportableConnectionOutcome(ConnectionOutcomeKind outcome) => outcome is
ConnectionOutcomeKind.Connected
or ConnectionOutcomeKind.Cancelled
or ConnectionOutcomeKind.TimedOut
or ConnectionOutcomeKind.StaleHost
or ConnectionOutcomeKind.TransportFailed
or ConnectionOutcomeKind.FallbackOffered
or ConnectionOutcomeKind.AttemptExpired
or ConnectionOutcomeKind.NoHostPresence
or ConnectionOutcomeKind.MediatorUnavailable
or ConnectionOutcomeKind.PunchTimedOut
or ConnectionOutcomeKind.DirectConnectTimedOut
or ConnectionOutcomeKind.HostRejected
or ConnectionOutcomeKind.TransportError
or ConnectionOutcomeKind.ManagerStopped
or ConnectionOutcomeKind.Disposed;
public static bool IsBuildVersionValid(string? value) => public static bool IsBuildVersionValid(string? value) =>
!string.IsNullOrWhiteSpace(value) !string.IsNullOrWhiteSpace(value)
&& IsUtf8LengthWithin(value, ContractLimits.BuildVersionMaxBytes); && IsUtf8LengthWithin(value, ContractLimits.BuildVersionMaxBytes);
@@ -5,9 +5,13 @@
<RootNamespace>FinalFactory.Rendezvous.Contracts</RootNamespace> <RootNamespace>FinalFactory.Rendezvous.Contracts</RootNamespace>
<IsPackable>true</IsPackable> <IsPackable>true</IsPackable>
<PackageId>FinalFactory.Rendezvous.Contracts</PackageId> <PackageId>FinalFactory.Rendezvous.Contracts</PackageId>
<PackageReadmeFile>README.md</PackageReadmeFile>
<Description>Versioned transport-neutral contracts for Final Factory Rendezvous.</Description> <Description>Versioned transport-neutral contracts for Final Factory Rendezvous.</Description>
<PackageTags>final-factory;multiplayer;contracts;godot</PackageTags>
</PropertyGroup> </PropertyGroup>
<ItemGroup> <ItemGroup>
<PackageReference Include="System.Text.Json" /> <PackageReference Include="System.Text.Json" />
<None Update="README.md" Pack="true" PackagePath="\" />
<None Include="../../CHANGELOG.md" Pack="true" PackagePath="\" Link="CHANGELOG.md" />
</ItemGroup> </ItemGroup>
</Project> </Project>
@@ -0,0 +1,33 @@
using System.Text.Json.Serialization;
namespace FinalFactory.Rendezvous.Contracts;
public sealed class ReportConnectionOutcomeRequest
{
[JsonRequired]
public int ContractVersion { get; set; } = ContractLimits.ContractVersion;
[JsonRequired]
public ConnectionOutcomeKind Outcome { get; set; }
public ConnectionElapsedBucket ElapsedBucket { get; set; }
[Obsolete("Use ElapsedBucket. Exact elapsed time is accepted only for v1 compatibility and is not retained.")]
[JsonIgnore(Condition = JsonIgnoreCondition.WhenWritingDefault)]
public int ElapsedMilliseconds { get; set; }
[Obsolete("Diagnostic codes are accepted only for v1 compatibility and are not retained.")]
public string? DiagnosticCode { get; set; }
}
public sealed class ReportConnectionOutcomeResponse
{
[JsonRequired]
public int ContractVersion { get; set; } = ContractLimits.ContractVersion;
[JsonRequired]
public bool Accepted { get; set; }
[JsonRequired]
public bool IsDuplicate { get; set; }
}
@@ -37,6 +37,9 @@ public sealed class CreateJoinAttemptResponse
[JsonRequired] [JsonRequired]
public string ClientPunchCapability { get; set; } = string.Empty; public string ClientPunchCapability { get; set; } = string.Empty;
[JsonRequired]
public string ConnectionTicketDigest { get; set; } = string.Empty;
[JsonRequired] [JsonRequired]
public DateTimeOffset ExpiresAt { get; set; } public DateTimeOffset ExpiresAt { get; set; }
public NetworkEndpoint? DedicatedFallback { get; set; } public NetworkEndpoint? DedicatedFallback { get; set; }
@@ -53,6 +56,12 @@ public sealed class HostJoinAttempt
[JsonRequired] [JsonRequired]
public string HostPunchCapability { get; set; } = string.Empty; public string HostPunchCapability { get; set; } = string.Empty;
[JsonRequired]
public string ConnectionTicketDigest { get; set; } = string.Empty;
[JsonRequired]
public bool IsCancelled { get; set; }
[JsonRequired] [JsonRequired]
public DateTimeOffset ExpiresAt { get; set; } public DateTimeOffset ExpiresAt { get; set; }
} }
@@ -67,26 +76,3 @@ public sealed class BrowseHostJoinAttemptsResponse
public string? NextCursor { get; set; } public string? NextCursor { get; set; }
} }
public sealed class ReportConnectionOutcomeRequest
{
[JsonRequired]
public int ContractVersion { get; set; } = ContractLimits.ContractVersion;
[JsonRequired]
public ConnectionOutcomeKind Outcome { get; set; }
[JsonRequired]
public int ElapsedMilliseconds { get; set; }
public string? DiagnosticCode { get; set; }
}
public sealed class ReportConnectionOutcomeResponse
{
[JsonRequired]
public int ContractVersion { get; set; } = ContractLimits.ContractVersion;
[JsonRequired]
public bool Accepted { get; set; }
}
@@ -39,6 +39,8 @@ public sealed class SessionListing
[JsonRequired] [JsonRequired]
public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal); public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal);
public NetworkEndpoint? DedicatedFallback { get; set; }
} }
public sealed class RegisterSessionRequest public sealed class RegisterSessionRequest
@@ -75,6 +77,8 @@ public sealed class RegisterSessionRequest
[JsonRequired] [JsonRequired]
public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal); public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal);
public NetworkEndpoint? DedicatedFallback { get; set; }
} }
public sealed class RegisterSessionResponse public sealed class RegisterSessionResponse
@@ -147,6 +151,8 @@ public sealed class UpdateSessionRequest
[JsonRequired] [JsonRequired]
public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal); public Dictionary<string, string> Metadata { get; set; } = new(StringComparer.Ordinal);
public NetworkEndpoint? DedicatedFallback { get; set; }
} }
public sealed class DeleteSessionRequest public sealed class DeleteSessionRequest
@@ -0,0 +1,8 @@
# FinalFactory.Rendezvous.Contracts
Transport-neutral v1 HTTP/UDP contract types for Final Factory Rendezvous.
The package targets `netstandard2.1`, contains no Godot or LiteNetLib dependency,
and is versioned with the Client package and server release.
Compatibility and migration policy is maintained in the repository's
`docs/releases/README.md` document.
@@ -25,7 +25,9 @@ public static class ContractJson
options.AllowTrailingCommas = false; options.AllowTrailingCommas = false;
options.DefaultIgnoreCondition = JsonIgnoreCondition.WhenWritingNull; options.DefaultIgnoreCondition = JsonIgnoreCondition.WhenWritingNull;
options.MaxDepth = 8; // Nine is the minimum that lets ASP.NET generate the nullable fallback
// OpenAPI schema; the 16 KiB HTTP body limit still bounds parser work.
options.MaxDepth = 9;
options.NumberHandling = JsonNumberHandling.Strict; options.NumberHandling = JsonNumberHandling.Strict;
options.PropertyNameCaseInsensitive = false; options.PropertyNameCaseInsensitive = false;
options.PropertyNamingPolicy = JsonNamingPolicy.CamelCase; options.PropertyNamingPolicy = JsonNamingPolicy.CamelCase;
@@ -0,0 +1,149 @@
using System.Security.Cryptography;
using System.Text;
namespace FinalFactory.Rendezvous.Contracts;
public sealed class NatIntroductionToken
{
public JoinAttemptId AttemptId { get; set; }
public string ConnectionTicket { get; set; } = string.Empty;
public override string ToString() =>
$"[NatIntroductionToken {AttemptId}; ticket redacted]";
}
public static class NatIntroductionTokenCodec
{
public const int EncodedLength = ContractLimits.DerivedCredentialCharacters;
private const int DecodedLength = 32;
private const int AttemptIdLength = 16;
private const int AuthenticatorLength = DecodedLength - AttemptIdLength;
public static string Encode(JoinAttemptId attemptId, string derivedAuthenticator)
{
if (attemptId.Value == Guid.Empty
|| !ContractValidation.IsConnectionTicketValid(derivedAuthenticator)
|| !TryDecodeBase64Url(derivedAuthenticator, out byte[]? authenticator)
|| authenticator.Length != DecodedLength)
{
throw new ArgumentException("The NAT introduction token fields are invalid.");
}
byte[] payload = new byte[DecodedLength];
try
{
if (!attemptId.Value.TryWriteBytes(payload.AsSpan(0, AttemptIdLength)))
{
throw new InvalidOperationException("The join attempt identifier could not be encoded.");
}
authenticator.AsSpan(0, AuthenticatorLength).CopyTo(payload.AsSpan(AttemptIdLength));
return EncodeBase64Url(payload);
}
finally
{
CryptographicOperations.ZeroMemory(authenticator);
CryptographicOperations.ZeroMemory(payload);
}
}
public static bool TryDecode(string? encoded, out NatIntroductionToken? token)
{
token = null;
if (!ContractValidation.IsConnectionTicketValid(encoded)
|| !TryDecodeBase64Url(encoded!, out byte[]? payload)
|| payload.Length != DecodedLength)
{
return false;
}
try
{
Guid attemptId = new(payload.AsSpan(0, AttemptIdLength));
if (attemptId == Guid.Empty)
{
return false;
}
token = new NatIntroductionToken
{
AttemptId = new JoinAttemptId(attemptId),
ConnectionTicket = encoded!,
};
return true;
}
finally
{
CryptographicOperations.ZeroMemory(payload);
}
}
public static string ComputeDigest(string connectionTicket)
{
if (!ContractValidation.IsConnectionTicketValid(connectionTicket))
{
throw new ArgumentException("The connection ticket is invalid.", nameof(connectionTicket));
}
byte[] encoded = Encoding.ASCII.GetBytes(connectionTicket);
byte[] digest;
using (SHA256 sha256 = SHA256.Create())
{
digest = sha256.ComputeHash(encoded);
}
CryptographicOperations.ZeroMemory(encoded);
try
{
return EncodeBase64Url(digest);
}
finally
{
CryptographicOperations.ZeroMemory(digest);
}
}
public static bool MatchesDigest(string? connectionTicket, string? expectedDigest)
{
if (!ContractValidation.IsConnectionTicketValid(connectionTicket)
|| !ContractValidation.IsConnectionTicketValid(expectedDigest))
{
return false;
}
byte[] actual = Encoding.ASCII.GetBytes(ComputeDigest(connectionTicket!));
byte[] expected = Encoding.ASCII.GetBytes(expectedDigest!);
try
{
return CryptographicOperations.FixedTimeEquals(actual, expected);
}
finally
{
CryptographicOperations.ZeroMemory(actual);
CryptographicOperations.ZeroMemory(expected);
}
}
private static bool TryDecodeBase64Url(string? encoded, out byte[] bytes)
{
bytes = [];
if (encoded is null || encoded.Length != EncodedLength)
{
return false;
}
try
{
bytes = Convert.FromBase64String(
encoded.Replace('-', '+').Replace('_', '/') + "=");
return true;
}
catch (FormatException)
{
return false;
}
}
private static string EncodeBase64Url(byte[] value) =>
Convert.ToBase64String(value).TrimEnd('=').Replace('+', '-').Replace('/', '_');
}
@@ -0,0 +1,131 @@
namespace FinalFactory.Rendezvous.Contracts;
public enum NatPunchPeerRole
{
HostPresence = 1,
Host = 2,
Client = 3,
}
public sealed class NatPunchRequestToken
{
public NatPunchPeerRole Role { get; set; }
public MediationHandle MediationHandle { get; set; }
public string Capability { get; set; } = string.Empty;
public override string ToString() => "[NatPunchRequestToken: capability redacted]";
}
public static class NatPunchRequestTokenCodec
{
public const int EncodedLength = ContractLimits.NatPunchRequestTokenCharacters;
private const string VersionPrefix = "rv1:";
private const int HandleLength = 32;
private const int CapabilityLength = ContractLimits.DerivedCredentialCharacters;
private const char Separator = ':';
private const char Padding = '.';
public static string Encode(
NatPunchPeerRole role,
MediationHandle mediationHandle,
string capability)
{
if (!TryGetRoleCode(role, out char roleCode)
|| mediationHandle.Value == Guid.Empty
|| capability is null
|| capability.Length != CapabilityLength
|| !ContractValidation.IsCapabilityValid(capability))
{
throw new ArgumentException("The NAT punch request token fields are invalid.");
}
string payload = string.Concat(
VersionPrefix,
roleCode,
Separator,
mediationHandle.Value.ToString("N"),
Separator,
capability);
return payload.PadRight(EncodedLength, Padding);
}
public static bool TryDecode(string? encoded, out NatPunchRequestToken? token)
{
token = null;
if (encoded is null
|| encoded.Length != EncodedLength
|| !encoded.StartsWith(VersionPrefix, StringComparison.Ordinal)
|| !TryParseRole(encoded[VersionPrefix.Length], out NatPunchPeerRole role))
{
return false;
}
int roleSeparator = VersionPrefix.Length + 1;
int handleOffset = roleSeparator + 1;
int capabilitySeparator = handleOffset + HandleLength;
int capabilityOffset = capabilitySeparator + 1;
int paddingOffset = capabilityOffset + CapabilityLength;
string handleText = encoded.Substring(handleOffset, HandleLength);
if (encoded[roleSeparator] != Separator
|| encoded[capabilitySeparator] != Separator
|| !Guid.TryParseExact(handleText, "N", out Guid handle)
|| handle == Guid.Empty
|| !string.Equals(handleText, handle.ToString("N"), StringComparison.Ordinal)
|| !ContainsOnlyPadding(encoded, paddingOffset))
{
return false;
}
string capability = encoded.Substring(capabilityOffset, CapabilityLength);
if (!ContractValidation.IsCapabilityValid(capability))
{
return false;
}
token = new NatPunchRequestToken
{
Role = role,
MediationHandle = new MediationHandle(handle),
Capability = capability,
};
return true;
}
private static bool TryGetRoleCode(NatPunchPeerRole role, out char code)
{
code = role switch
{
NatPunchPeerRole.HostPresence => 'p',
NatPunchPeerRole.Host => 'h',
NatPunchPeerRole.Client => 'c',
_ => default,
};
return code != default;
}
private static bool TryParseRole(char code, out NatPunchPeerRole role)
{
role = code switch
{
'p' => NatPunchPeerRole.HostPresence,
'h' => NatPunchPeerRole.Host,
'c' => NatPunchPeerRole.Client,
_ => default,
};
return role != default;
}
private static bool ContainsOnlyPadding(string value, int offset)
{
for (int index = offset; index < value.Length; index++)
{
if (value[index] != Padding)
{
return false;
}
}
return true;
}
}
@@ -0,0 +1,111 @@
using System.ComponentModel.DataAnnotations;
namespace FinalFactory.Rendezvous.Server.Abuse;
internal sealed class AbuseProtectionOptions
{
public const string SectionName = "Rendezvous:AbuseProtection";
[Range(1, 60)]
public int WindowSeconds { get; set; } = 1;
[Range(1_000, 1_000_000)]
public int MaxTrackedKeys { get; set; } = 100_000;
[Range(0, 100_000)]
public int CriticalTrackedKeyReserve { get; set; } = 2_048;
[Range(1_000, 999_999)]
public int UdpTrackedKeyLimit { get; set; } = 70_000;
public string[] TrustedProxyAddresses { get; set; } = [];
public string[] OperatorAllowedAddresses { get; set; } = [];
[Range(1, 100_000)]
public int HealthGlobalRequestsPerWindow { get; set; } = 1_000;
[Range(1, 10_000)]
public int HealthGlobalConcurrency { get; set; } = 32;
[Range(1, 100_000)]
public int HealthIpPrefixRequestsPerWindow { get; set; } = 120;
[Range(1, 1_000)]
public int HealthIpPrefixConcurrency { get; set; } = 8;
[Range(1, 100_000)]
public int OperatorGlobalRequestsPerWindow { get; set; } = 1_000;
[Range(1, 10_000)]
public int OperatorGlobalConcurrency { get; set; } = 32;
[Range(1, 100_000)]
public int OperatorIpPrefixRequestsPerWindow { get; set; } = 120;
[Range(1, 1_000)]
public int OperatorIpPrefixConcurrency { get; set; } = 8;
[Range(1, 1_000_000)]
public int HttpGlobalRequestsPerWindow { get; set; } = 20_000;
[Range(1, 1_000_000)]
public int HttpOptionalRequestsPerWindow { get; set; } = 18_000;
[Range(1, 100_000)]
public int HttpIpPrefixRequestsPerWindow { get; set; } = 500;
[Range(1, 100_000)]
public int HttpOptionalIpPrefixRequestsPerWindow { get; set; } = 450;
[Range(1, 1_000_000)]
public int HttpOperationRequestsPerWindow { get; set; } = 5_000;
[Range(1, 1_000_000)]
public int HttpTenantRequestsPerWindow { get; set; } = 2_000;
[Range(1, 100_000)]
public int HttpPrincipalRequestsPerWindow { get; set; } = 500;
[Range(1, 100_000)]
public int HttpResourceRequestsPerWindow { get; set; } = 200;
[Range(1, 100_000)]
public int HttpGlobalConcurrency { get; set; } = 1_024;
[Range(1, 100_000)]
public int HttpOptionalConcurrency { get; set; } = 768;
[Range(1, 10_000)]
public int HttpIpPrefixConcurrency { get; set; } = 64;
[Range(1, 10_000)]
public int HttpOptionalIpPrefixConcurrency { get; set; } = 48;
[Range(1, 100_000)]
public int HttpOperationConcurrency { get; set; } = 256;
[Range(1, 100_000)]
public int HttpTenantConcurrency { get; set; } = 256;
[Range(1, 10_000)]
public int HttpPrincipalConcurrency { get; set; } = 32;
[Range(1, 10_000)]
public int HttpResourceConcurrency { get; set; } = 16;
[Range(1, 10_000_000)]
public int UdpGlobalDatagramsPerWindow { get; set; } = 100_000;
[Range(1, 1_000_000)]
public int UdpIpPrefixDatagramsPerWindow { get; set; } = 2_000;
[Range(1, 10_000_000)]
public int UdpOperationDatagramsPerWindow { get; set; } = 50_000;
[Range(1, 100_000)]
public int UdpCapabilityDatagramsPerWindow { get; set; } = 120;
[Range(1, 100_000)]
public int UdpResourceDatagramsPerWindow { get; set; } = 240;
}
@@ -0,0 +1,530 @@
using System.Buffers;
using System.Net;
using System.Security.Cryptography;
using System.Text;
using FinalFactory.Rendezvous.Server.Observability;
using Microsoft.Extensions.Options;
namespace FinalFactory.Rendezvous.Server.Abuse;
internal sealed class AbuseProtectionService
{
private readonly AbuseProtectionOptions _options;
private readonly TimeProvider _timeProvider;
private readonly TrackerState _httpTracker;
private readonly TrackerState _udpTracker;
private readonly RendezvousTelemetry? _telemetry;
private readonly HashSet<string> _operatorAllowedAddresses;
public AbuseProtectionService(
IOptions<AbuseProtectionOptions> options,
TimeProvider? timeProvider = null,
RendezvousTelemetry? telemetry = null)
{
_options = options.Value;
_timeProvider = timeProvider ?? TimeProvider.System;
_telemetry = telemetry;
_operatorAllowedAddresses = options.Value.OperatorAllowedAddresses
.Select(static value => IPAddress.TryParse(value, out IPAddress? address)
? NormalizeAddress(address).ToString()
: string.Empty)
.Where(static value => value.Length > 0)
.ToHashSet(StringComparer.Ordinal);
DateTimeOffset now = _timeProvider.GetUtcNow();
_httpTracker = new(now);
_udpTracker = new(now);
}
public bool TryAcquireHttpIngress(
IPAddress? remoteAddress,
string operation,
out AbuseLease? lease,
out int retryAfterSeconds)
{
string prefix = GetNetworkPrefix(remoteAddress);
List<RateDimension> rates =
[
new("http:rate:global", _options.HttpGlobalRequestsPerWindow),
new($"http:rate:ip:{prefix}", _options.HttpIpPrefixRequestsPerWindow),
new($"http:rate:operation:{operation}", _options.HttpOperationRequestsPerWindow),
];
List<RateDimension> concurrency =
[
new("http:concurrency:global", _options.HttpGlobalConcurrency),
new($"http:concurrency:ip:{prefix}", _options.HttpIpPrefixConcurrency),
new($"http:concurrency:operation:{operation}", _options.HttpOperationConcurrency),
];
if (!IsLeaseCriticalOperation(operation))
{
rates.Add(new("http:rate:optional", _options.HttpOptionalRequestsPerWindow));
rates.Add(new($"http:rate:optional-ip:{prefix}",
_options.HttpOptionalIpPrefixRequestsPerWindow));
concurrency.Add(new("http:concurrency:optional", _options.HttpOptionalConcurrency));
concurrency.Add(new($"http:concurrency:optional-ip:{prefix}",
_options.HttpOptionalIpPrefixConcurrency));
}
return TryAcquire(
[.. rates],
[.. concurrency],
TrackerDomain.Http,
IsLeaseCriticalOperation(operation),
out lease,
out retryAfterSeconds);
}
public bool TryAcquireHealthIngress(
IPAddress? remoteAddress,
out AbuseLease? lease,
out int retryAfterSeconds)
{
string prefix = GetNetworkPrefix(remoteAddress);
RateDimension[] rates =
[
new("health:rate:global", _options.HealthGlobalRequestsPerWindow),
new($"health:rate:ip:{prefix}", _options.HealthIpPrefixRequestsPerWindow),
];
RateDimension[] concurrency =
[
new("health:concurrency:global", _options.HealthGlobalConcurrency),
new($"health:concurrency:ip:{prefix}", _options.HealthIpPrefixConcurrency),
];
return TryAcquire(
rates,
concurrency,
TrackerDomain.Http,
true,
out lease,
out retryAfterSeconds);
}
public bool IsOperatorSourceAllowed(IPAddress? remoteAddress) =>
remoteAddress is not null
&& _operatorAllowedAddresses.Contains(NormalizeAddress(remoteAddress).ToString());
public bool TryAcquireOperatorIngress(
IPAddress? remoteAddress,
out AbuseLease? lease,
out int retryAfterSeconds)
{
string prefix = GetNetworkPrefix(remoteAddress);
RateDimension[] rates =
[
new("operator:rate:global", _options.OperatorGlobalRequestsPerWindow),
new($"operator:rate:ip:{prefix}", _options.OperatorIpPrefixRequestsPerWindow),
];
RateDimension[] concurrency =
[
new("operator:concurrency:global", _options.OperatorGlobalConcurrency),
new($"operator:concurrency:ip:{prefix}", _options.OperatorIpPrefixConcurrency),
];
return TryAcquire(
rates,
concurrency,
TrackerDomain.Http,
true,
out lease,
out retryAfterSeconds);
}
public bool TryAcquireHttpIdentity(
string operation,
string? tenant,
string? principal,
string? resource,
out AbuseLease? lease,
out int retryAfterSeconds) => TryAcquireHttpIdentity(
operation,
null,
tenant,
principal,
resource,
out lease,
out retryAfterSeconds);
public bool TryAcquireHttpIdentity(
string operation,
IPAddress? remoteAddress,
string? tenant,
string? principal,
string? resource,
out AbuseLease? lease,
out int retryAfterSeconds)
{
string sourcePrefix = GetNetworkPrefix(remoteAddress);
List<RateDimension> rates = [];
List<RateDimension> concurrency = [];
AddDimension(rates, concurrency, "tenant", tenant,
_options.HttpTenantRequestsPerWindow, _options.HttpTenantConcurrency);
AddDimension(rates, concurrency, "principal", principal,
_options.HttpPrincipalRequestsPerWindow, _options.HttpPrincipalConcurrency);
AddDimension(rates, concurrency, "resource", resource,
_options.HttpResourceRequestsPerWindow, _options.HttpResourceConcurrency);
return TryAcquire(
[.. rates],
[.. concurrency],
TrackerDomain.Http,
IsLeaseCriticalOperation(operation),
out lease,
out retryAfterSeconds);
void AddDimension(
List<RateDimension> rateDimensions,
List<RateDimension> concurrencyDimensions,
string kind,
string? value,
int rateLimit,
int concurrencyLimit)
{
if (string.IsNullOrEmpty(value))
{
return;
}
if (kind == "resource")
{
string validationKey =
$"http:source-resource:{operation}:{sourcePrefix}:{value}";
rateDimensions.Add(new($"{validationKey}:rate", rateLimit));
concurrencyDimensions.Add(new($"{validationKey}:concurrency", concurrencyLimit));
}
string key = kind == "resource"
? $"http:resource-scoped:{operation}:{tenant ?? string.Empty}|{principal ?? string.Empty}:{value}"
: $"http:{kind}:{operation}:{value}";
rateDimensions.Add(new($"{key}:rate", rateLimit));
concurrencyDimensions.Add(new($"{key}:concurrency", concurrencyLimit));
}
}
public bool TryAcceptUdpIngress(IPAddress? remoteAddress, string operation)
{
string prefix = GetNetworkPrefix(remoteAddress);
RateDimension[] rates =
[
new("udp:rate:global", _options.UdpGlobalDatagramsPerWindow),
new($"udp:rate:ip:{prefix}", _options.UdpIpPrefixDatagramsPerWindow),
new($"udp:rate:operation:{operation}", _options.UdpOperationDatagramsPerWindow),
];
return TryAcquire(
rates,
[],
TrackerDomain.Udp,
false,
out AbuseLease? lease,
out _)
&& DisposeAccepted(lease);
}
public bool TryAcceptUdpIdentity(
string operation,
string capability,
string resource) => TryAcceptUdpIdentity(
operation,
null,
capability,
resource);
public bool TryAcceptUdpIdentity(
string operation,
IPAddress? remoteAddress,
string capability,
string resource)
{
string sourcePrefix = GetNetworkPrefix(remoteAddress);
string capabilityFingerprint = FingerprintSecret(capability);
RateDimension[] rates =
[
new($"udp:rate:capability:{operation}:{capabilityFingerprint}",
_options.UdpCapabilityDatagramsPerWindow),
new($"udp:rate:source-resource:{operation}:{sourcePrefix}:{resource}",
_options.UdpResourceDatagramsPerWindow),
new($"udp:rate:resource:{operation}:{capabilityFingerprint}:{resource}",
_options.UdpResourceDatagramsPerWindow),
];
return TryAcquire(
rates,
[],
TrackerDomain.Udp,
false,
out AbuseLease? lease,
out _)
&& DisposeAccepted(lease);
}
public static string FingerprintSecret(string secret)
{
int byteCount = Encoding.UTF8.GetByteCount(secret);
byte[]? rented = null;
Span<byte> encoded = byteCount <= 1_024
? stackalloc byte[byteCount]
: (rented = ArrayPool<byte>.Shared.Rent(byteCount)).AsSpan(0, byteCount);
Span<byte> digest = stackalloc byte[32];
try
{
_ = Encoding.UTF8.GetBytes(secret, encoded);
_ = SHA256.HashData(encoded, digest);
return Convert.ToHexString(digest[..12]);
}
finally
{
CryptographicOperations.ZeroMemory(encoded);
CryptographicOperations.ZeroMemory(digest);
if (rented is not null)
{
ArrayPool<byte>.Shared.Return(rented);
}
}
}
internal int TrackedKeyCount
{
get
{
int http;
int udp;
lock (_httpTracker.Gate)
{
http = _httpTracker.WindowCounts.Count + _httpTracker.ConcurrencyCounts.Count;
}
lock (_udpTracker.Gate)
{
udp = _udpTracker.WindowCounts.Count + _udpTracker.ConcurrencyCounts.Count;
}
return http + udp;
}
}
private bool TryAcquire(
ReadOnlySpan<RateDimension> rates,
ReadOnlySpan<RateDimension> concurrency,
TrackerDomain domain,
bool canUseCriticalReserve,
out AbuseLease? lease,
out int retryAfterSeconds)
{
TrackerState tracker = domain == TrackerDomain.Udp ? _udpTracker : _httpTracker;
bool accepted;
lock (tracker.Gate)
{
accepted = TryAcquireLocked(
tracker,
rates,
concurrency,
domain,
canUseCriticalReserve,
out lease,
out retryAfterSeconds);
}
if (!accepted)
{
_telemetry?.RecordLimiterDrop(
domain == TrackerDomain.Udp ? "udp" : "http",
"rate-or-concurrency");
}
return accepted;
}
private bool TryAcquireLocked(
TrackerState tracker,
ReadOnlySpan<RateDimension> rates,
ReadOnlySpan<RateDimension> concurrency,
TrackerDomain domain,
bool canUseCriticalReserve,
out AbuseLease? lease,
out int retryAfterSeconds)
{
DateTimeOffset now = _timeProvider.GetUtcNow();
TimeSpan window = TimeSpan.FromSeconds(_options.WindowSeconds);
if (now - tracker.WindowStartedAt >= window || now < tracker.WindowStartedAt)
{
tracker.WindowCounts.Clear();
tracker.WindowStartedAt = now;
}
retryAfterSeconds = Math.Max(
1,
(int)Math.Ceiling((window - (now - tracker.WindowStartedAt)).TotalSeconds));
int stagedNewKeys = 0;
int partitionLimit = domain == TrackerDomain.Udp
? _options.UdpTrackedKeyLimit
: _options.MaxTrackedKeys - _options.UdpTrackedKeyLimit;
int maxTrackedKeys = domain == TrackerDomain.Udp || canUseCriticalReserve
? partitionLimit
: partitionLimit - _options.CriticalTrackedKeyReserve;
if (!CanAcquireAll(
tracker,
tracker.WindowCounts,
rates,
maxTrackedKeys,
ref stagedNewKeys)
|| !CanAcquireAll(
tracker,
tracker.ConcurrencyCounts,
concurrency,
maxTrackedKeys,
ref stagedNewKeys))
{
lease = null;
return false;
}
foreach (RateDimension dimension in rates)
{
tracker.WindowCounts[dimension.Key] =
tracker.WindowCounts.GetValueOrDefault(dimension.Key) + 1;
}
if (concurrency.IsEmpty)
{
lease = null;
return true;
}
string[] acquiredConcurrency = new string[concurrency.Length];
for (int index = 0; index < concurrency.Length; index++)
{
RateDimension dimension = concurrency[index];
tracker.ConcurrencyCounts[dimension.Key] =
tracker.ConcurrencyCounts.GetValueOrDefault(dimension.Key) + 1;
acquiredConcurrency[index] = dimension.Key;
}
lease = new AbuseLease(this, tracker, acquiredConcurrency);
return true;
}
private static bool CanAcquireAll(
TrackerState tracker,
Dictionary<string, int> counts,
ReadOnlySpan<RateDimension> dimensions,
int maxTrackedKeys,
ref int stagedNewKeys)
{
foreach (RateDimension dimension in dimensions)
{
if (counts.TryGetValue(dimension.Key, out int current))
{
if (current >= dimension.Limit)
{
return false;
}
continue;
}
stagedNewKeys++;
if (tracker.WindowCounts.Count + tracker.ConcurrencyCounts.Count + stagedNewKeys
> maxTrackedKeys)
{
return false;
}
}
return true;
}
private static void Release(TrackerState tracker, string[] keys)
{
lock (tracker.Gate)
{
foreach (string key in keys)
{
if (!tracker.ConcurrencyCounts.TryGetValue(key, out int current))
{
continue;
}
if (current <= 1)
{
tracker.ConcurrencyCounts.Remove(key);
}
else
{
tracker.ConcurrencyCounts[key] = current - 1;
}
}
}
}
private static bool DisposeAccepted(AbuseLease? lease)
{
lease?.Dispose();
return true;
}
private static bool IsLeaseCriticalOperation(string operation) => operation is
"RenewSessionLease" or "UpdateSession" or "DeleteSession";
private static string GetNetworkPrefix(IPAddress? address)
{
if (address is null)
{
return "unknown";
}
IPAddress normalized = address.IsIPv4MappedToIPv6 ? address.MapToIPv4() : address;
byte[] bytes = normalized.GetAddressBytes();
if (bytes.Length == 4)
{
bytes[3] = 0;
return $"4:{Convert.ToHexString(bytes)}:24";
}
if (bytes.Length == 16)
{
Array.Clear(bytes, 7, 9);
return $"6:{Convert.ToHexString(bytes)}:56";
}
return "unknown";
}
private static IPAddress NormalizeAddress(IPAddress address) =>
address.IsIPv4MappedToIPv6 ? address.MapToIPv4() : address;
private readonly record struct RateDimension(string Key, int Limit);
private enum TrackerDomain
{
Http,
Udp,
}
internal sealed class TrackerState(DateTimeOffset windowStartedAt)
{
public object Gate { get; } = new();
public Dictionary<string, int> WindowCounts { get; } = new(StringComparer.Ordinal);
public Dictionary<string, int> ConcurrencyCounts { get; } = new(StringComparer.Ordinal);
public DateTimeOffset WindowStartedAt { get; set; } = windowStartedAt;
}
internal sealed class AbuseLease : IDisposable
{
private AbuseProtectionService? _owner;
private readonly TrackerState _tracker;
private readonly string[] _keys;
internal AbuseLease(
AbuseProtectionService owner,
TrackerState tracker,
string[] keys)
{
_owner = owner;
_tracker = tracker;
_keys = keys;
}
public void Dispose()
{
if (Interlocked.Exchange(ref _owner, null) is not null)
{
Release(_tracker, _keys);
}
}
}
}
@@ -0,0 +1,135 @@
using FinalFactory.Rendezvous.Contracts;
using Microsoft.AspNetCore.Http.Features;
namespace FinalFactory.Rendezvous.Server.Abuse;
internal sealed class HttpAbuseProtectionMiddleware(
RequestDelegate next,
AbuseProtectionService protection)
{
public async Task InvokeAsync(HttpContext context)
{
IHttpMaxRequestBodySizeFeature? bodySize =
context.Features.Get<IHttpMaxRequestBodySizeFeature>();
if (bodySize is { IsReadOnly: false })
{
bodySize.MaxRequestBodySize = ContractLimits.HttpRequestMaxBytes;
}
string operation = context.GetEndpoint()?.Metadata.GetMetadata<IEndpointNameMetadata>()
?.EndpointName ?? "Unmatched";
bool healthEndpoint = operation is "GetLiveness" or "GetReadiness";
bool operatorEndpoint = operation is
"GetOperatorStatus"
or "RevokeOperatorListing"
or "RevokeOperatorPrincipal"
or "RevokeOperatorSigningKey"
or "BeginOperatorDrain";
if (operatorEndpoint
&& !protection.IsOperatorSourceAllowed(context.Connection.RemoteIpAddress))
{
bool deniedSourceAdmitted = protection.TryAcquireHttpIngress(
context.Connection.RemoteIpAddress,
"Unmatched",
out AbuseProtectionService.AbuseLease? deniedSourceLease,
out int deniedRetryAfterSeconds);
using (deniedSourceLease)
{
if (!deniedSourceAdmitted)
{
context.Response.Headers.RetryAfter = deniedRetryAfterSeconds.ToString(
System.Globalization.CultureInfo.InvariantCulture);
await WriteErrorAsync(
context,
StatusCodes.Status429TooManyRequests,
RendezvousErrorCode.RateLimited,
"The request rate limit was exceeded.",
deniedRetryAfterSeconds).ConfigureAwait(false);
return;
}
await WriteErrorAsync(
context,
StatusCodes.Status404NotFound,
RendezvousErrorCode.NotFound,
"The requested resource was not found.").ConfigureAwait(false);
}
return;
}
AbuseProtectionService.AbuseLease? lease;
int retryAfterSeconds;
bool acquired;
if (healthEndpoint)
{
acquired = protection.TryAcquireHealthIngress(
context.Connection.RemoteIpAddress,
out lease,
out retryAfterSeconds);
}
else if (operatorEndpoint)
{
acquired = protection.TryAcquireOperatorIngress(
context.Connection.RemoteIpAddress,
out lease,
out retryAfterSeconds);
}
else
{
acquired = protection.TryAcquireHttpIngress(
context.Connection.RemoteIpAddress,
operation,
out lease,
out retryAfterSeconds);
}
if (!acquired)
{
context.Response.Headers.RetryAfter = retryAfterSeconds.ToString(
System.Globalization.CultureInfo.InvariantCulture);
await WriteErrorAsync(
context,
StatusCodes.Status429TooManyRequests,
RendezvousErrorCode.RateLimited,
"The request rate limit was exceeded.",
retryAfterSeconds).ConfigureAwait(false);
return;
}
using (lease)
{
if (context.Request.ContentLength > ContractLimits.HttpRequestMaxBytes)
{
await WriteErrorAsync(
context,
StatusCodes.Status413PayloadTooLarge,
RendezvousErrorCode.InvalidRequest,
"The request body exceeds the supported size.").ConfigureAwait(false);
return;
}
await next(context).ConfigureAwait(false);
}
}
private static Task WriteErrorAsync(
HttpContext context,
int status,
RendezvousErrorCode code,
string message,
int? retryAfterSeconds = null)
{
context.Response.StatusCode = status;
return context.Response.WriteAsJsonAsync(
new ApiError
{
Code = code,
Message = message,
RetryAfterSeconds = retryAfterSeconds,
},
ContractJson.Options,
contentType: "application/json",
cancellationToken: context.RequestAborted);
}
}
@@ -0,0 +1,24 @@
using System.Net;
using Microsoft.AspNetCore.HttpOverrides;
namespace FinalFactory.Rendezvous.Server.Abuse;
internal static class TrustedProxyForwarding
{
public static bool IsEnabled(AbuseProtectionOptions options) =>
options.TrustedProxyAddresses is { Length: > 0 };
public static void Configure(
ForwardedHeadersOptions forwarded,
AbuseProtectionOptions abuse)
{
forwarded.ForwardedHeaders = ForwardedHeaders.XForwardedFor;
forwarded.ForwardLimit = 1;
forwarded.KnownProxies.Clear();
forwarded.KnownIPNetworks.Clear();
foreach (string address in abuse.TrustedProxyAddresses ?? [])
{
forwarded.KnownProxies.Add(IPAddress.Parse(address));
}
}
}
@@ -0,0 +1,103 @@
using System.Security.Cryptography;
using System.Text;
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Server.Browser;
internal sealed class EphemeralCursorProtector : IDisposable
{
private readonly byte[] _key = RandomNumberGenerator.GetBytes(32);
private bool _disposed;
public string Protect(string prefix, ReadOnlySpan<byte> payload)
{
ObjectDisposedException.ThrowIf(_disposed, this);
string content = $"{prefix}.{EncodeBytes(payload)}";
byte[] signature = HMACSHA256.HashData(_key, Encoding.ASCII.GetBytes(content));
try
{
string cursor = $"{content}.{EncodeBytes(signature)}";
return ContractValidation.IsCursorValid(cursor)
? cursor
: throw new InvalidOperationException("The protected cursor exceeds its contract limit.");
}
finally
{
CryptographicOperations.ZeroMemory(signature);
}
}
public bool TryUnprotect(string prefix, string? cursor, out byte[] payload)
{
payload = [];
if (_disposed || !ContractValidation.IsCursorValid(cursor))
{
return false;
}
string[] segments = cursor!.Split('.');
if (segments.Length != 3 || !string.Equals(segments[0], prefix, StringComparison.Ordinal))
{
return false;
}
byte[] expected = HMACSHA256.HashData(
_key,
Encoding.ASCII.GetBytes($"{segments[0]}.{segments[1]}"));
if (!TryDecodeBytes(segments[2], out byte[] supplied))
{
CryptographicOperations.ZeroMemory(expected);
return false;
}
bool validSignature = supplied.Length == expected.Length
&& CryptographicOperations.FixedTimeEquals(supplied, expected);
CryptographicOperations.ZeroMemory(supplied);
CryptographicOperations.ZeroMemory(expected);
return validSignature && TryDecodeBytes(segments[1], out payload);
}
public void Dispose()
{
if (!_disposed)
{
_disposed = true;
CryptographicOperations.ZeroMemory(_key);
}
}
public override string ToString() => "[EphemeralCursorProtector: key redacted]";
private static string EncodeBytes(ReadOnlySpan<byte> bytes) => Convert
.ToBase64String(bytes)
.TrimEnd('=')
.Replace('+', '-')
.Replace('/', '_');
private static bool TryDecodeBytes(string value, out byte[] bytes)
{
bytes = [];
if (string.IsNullOrEmpty(value)
|| value.Any(static character =>
character is not (>= 'A' and <= 'Z')
and not (>= 'a' and <= 'z')
and not (>= '0' and <= '9')
and not '-'
and not '_'))
{
return false;
}
string padded = value.Replace('-', '+').Replace('_', '/');
padded += (padded.Length % 4) switch { 0 => "", 2 => "==", 3 => "=", _ => "!" };
try
{
bytes = Convert.FromBase64String(padded);
return true;
}
catch (FormatException)
{
return false;
}
}
}
@@ -1,5 +1,4 @@
using System.Security.Cryptography; using System.Security.Cryptography;
using System.Text;
using System.Text.Json; using System.Text.Json;
using System.Text.Json.Serialization; using System.Text.Json.Serialization;
using FinalFactory.Rendezvous.Contracts; using FinalFactory.Rendezvous.Contracts;
@@ -10,12 +9,10 @@ namespace FinalFactory.Rendezvous.Server.Browser;
internal sealed class SessionBrowserCursorCodec : IDisposable internal sealed class SessionBrowserCursorCodec : IDisposable
{ {
private const string Prefix = "rvc1"; private const string Prefix = "rvc1";
private readonly byte[] _key = RandomNumberGenerator.GetBytes(32); private readonly EphemeralCursorProtector _protector = new();
private bool _disposed;
public string Encode(VisibleListingQuery query, SessionListingId after, DateTimeOffset now) public string Encode(VisibleListingQuery query, SessionListingId after, DateTimeOffset now)
{ {
ObjectDisposedException.ThrowIf(_disposed, this);
BrowserCursorPayload payload = new() BrowserCursorPayload payload = new()
{ {
GameId = query.Scope.GameId.Value, GameId = query.Scope.GameId.Value,
@@ -26,19 +23,14 @@ internal sealed class SessionBrowserCursorCodec : IDisposable
AfterListingId = after.ToString(), AfterListingId = after.ToString(),
ExpiresAtUnixSeconds = now.AddMinutes(5).ToUnixTimeSeconds(), ExpiresAtUnixSeconds = now.AddMinutes(5).ToUnixTimeSeconds(),
}; };
string encoded = EncodeBytes(JsonSerializer.SerializeToUtf8Bytes(payload, ContractJson.Options)); byte[] encoded = JsonSerializer.SerializeToUtf8Bytes(payload, ContractJson.Options);
string content = $"{Prefix}.{encoded}";
byte[] signature = HMACSHA256.HashData(_key, Encoding.ASCII.GetBytes(content));
try try
{ {
string cursor = $"{content}.{EncodeBytes(signature)}"; return _protector.Protect(Prefix, encoded);
return ContractValidation.IsCursorValid(cursor)
? cursor
: throw new InvalidOperationException("The browser cursor exceeds its contract limit.");
} }
finally finally
{ {
CryptographicOperations.ZeroMemory(signature); CryptographicOperations.ZeroMemory(encoded);
} }
} }
@@ -57,31 +49,7 @@ internal sealed class SessionBrowserCursorCodec : IDisposable
return true; return true;
} }
if (_disposed || !ContractValidation.IsCursorValid(cursor)) if (!_protector.TryUnprotect(Prefix, cursor, out byte[] encodedPayload))
{
return false;
}
string[] segments = cursor.Split('.');
if (segments.Length != 3 || !string.Equals(segments[0], Prefix, StringComparison.Ordinal))
{
return false;
}
byte[] expected = HMACSHA256.HashData(
_key,
Encoding.ASCII.GetBytes($"{segments[0]}.{segments[1]}"));
if (!TryDecodeBytes(segments[2], out byte[] supplied))
{
CryptographicOperations.ZeroMemory(expected);
return false;
}
bool validSignature = supplied.Length == expected.Length
&& CryptographicOperations.FixedTimeEquals(supplied, expected);
CryptographicOperations.ZeroMemory(supplied);
CryptographicOperations.ZeroMemory(expected);
if (!validSignature || !TryDecodeBytes(segments[1], out byte[] encodedPayload))
{ {
return false; return false;
} }
@@ -118,64 +86,30 @@ internal sealed class SessionBrowserCursorCodec : IDisposable
return true; return true;
} }
public void Dispose() public void Dispose() => _protector.Dispose();
{
if (!_disposed)
{
_disposed = true;
CryptographicOperations.ZeroMemory(_key);
}
}
public override string ToString() => "[SessionBrowserCursorCodec: key and cursors redacted]"; public override string ToString() => "[SessionBrowserCursorCodec: key and cursors redacted]";
private static string EncodeBytes(ReadOnlySpan<byte> bytes) => Convert
.ToBase64String(bytes)
.TrimEnd('=')
.Replace('+', '-')
.Replace('/', '_');
private static bool TryDecodeBytes(string value, out byte[] bytes)
{
bytes = [];
if (string.IsNullOrEmpty(value)
|| value.Any(static character =>
character is not (>= 'A' and <= 'Z')
and not (>= 'a' and <= 'z')
and not (>= '0' and <= '9')
and not '-'
and not '_'))
{
return false;
}
string padded = value.Replace('-', '+').Replace('_', '/');
padded += (padded.Length % 4) switch { 0 => "", 2 => "==", 3 => "=", _ => "!" };
try
{
bytes = Convert.FromBase64String(padded);
return true;
}
catch (FormatException)
{
return false;
}
}
} }
internal sealed class BrowserCursorPayload internal sealed class BrowserCursorPayload
{ {
[JsonRequired] [JsonRequired]
public string GameId { get; set; } = string.Empty; public string GameId { get; set; } = string.Empty;
[JsonRequired] [JsonRequired]
public string EnvironmentId { get; set; } = string.Empty; public string EnvironmentId { get; set; } = string.Empty;
[JsonRequired] [JsonRequired]
public uint ProtocolVersion { get; set; } public uint ProtocolVersion { get; set; }
public string? RegionId { get; set; } public string? RegionId { get; set; }
[JsonRequired] [JsonRequired]
public bool ExcludeFull { get; set; } public bool ExcludeFull { get; set; }
[JsonRequired] [JsonRequired]
public string AfterListingId { get; set; } = string.Empty; public string AfterListingId { get; set; } = string.Empty;
[JsonRequired] [JsonRequired]
public long ExpiresAtUnixSeconds { get; set; } public long ExpiresAtUnixSeconds { get; set; }
} }
@@ -153,5 +153,6 @@ internal sealed class SessionBrowserService(
static item => item.Key, static item => item.Key,
static item => item.Value, static item => item.Value,
StringComparer.Ordinal), StringComparer.Ordinal),
DedicatedFallback = StoredListing.CopyEndpoint(stored.Definition.DedicatedFallback),
}; };
} }
@@ -0,0 +1,141 @@
using FinalFactory.Rendezvous.Contracts;
using FinalFactory.Rendezvous.Server.Observability;
using FinalFactory.Rendezvous.Server.Sessions;
using FinalFactory.Rendezvous.Server.State;
namespace FinalFactory.Rendezvous.Server.ConnectionOutcomes;
internal sealed record ConnectionOutcomeServiceResult(
RendezvousErrorCode Error,
ReportConnectionOutcomeResponse? Value = null)
{
public bool Succeeded => Error == RendezvousErrorCode.None;
}
internal sealed class ConnectionOutcomeMetrics
{
private readonly object _gate = new();
private readonly Dictionary<(ConnectionOutcomeKind, ConnectionElapsedBucket), long> _counts = [];
private readonly RendezvousTelemetry? _telemetry;
public ConnectionOutcomeMetrics(RendezvousTelemetry? telemetry = null) =>
_telemetry = telemetry;
internal void Record(ConnectionOutcomeKind outcome, ConnectionElapsedBucket elapsedBucket)
{
lock (_gate)
{
(ConnectionOutcomeKind, ConnectionElapsedBucket) key = (outcome, elapsedBucket);
_counts.TryGetValue(key, out long count);
_counts[key] = count + 1;
}
_telemetry?.RecordConnectionOutcome(outcome.ToString(), elapsedBucket.ToString());
}
internal long GetCount(ConnectionOutcomeKind outcome, ConnectionElapsedBucket elapsedBucket)
{
lock (_gate)
{
return _counts.GetValueOrDefault((outcome, elapsedBucket));
}
}
}
internal sealed class ConnectionOutcomeService(
IEphemeralRendezvousStore store,
ISessionCapabilityService capabilities,
ConnectionOutcomeMetrics metrics)
{
internal ConnectionOutcomeServiceResult Report(
JoinAttemptId attemptId,
string? clientPunchCapability,
ReportConnectionOutcomeRequest request,
CancellationToken cancellationToken = default)
{
ArgumentNullException.ThrowIfNull(request);
RendezvousErrorCode version = ContractValidation.ValidateContractVersion(
request.ContractVersion);
if (version != RendezvousErrorCode.None)
{
return new(version);
}
if (attemptId.Value == Guid.Empty
|| !ContractValidation.IsCapabilityValid(clientPunchCapability)
|| !TryNormalizeReport(request, out ConnectionOutcomeKind outcome, out ConnectionElapsedBucket elapsedBucket)
|| !capabilities.TryFingerprint(
clientPunchCapability,
out SecretFingerprint capabilityFingerprint))
{
return new(RendezvousErrorCode.InvalidRequest);
}
StoreResult<StoredConnectionOutcome> reported = store.ReportConnectionOutcome(new(
attemptId,
capabilityFingerprint,
outcome,
elapsedBucket), cancellationToken);
if (!reported.Succeeded)
{
return new(reported.Code.ToContractError());
}
if (!reported.IsIdempotentReplay)
{
metrics.Record(outcome, elapsedBucket);
}
return new(RendezvousErrorCode.None, new ReportConnectionOutcomeResponse
{
Accepted = true,
IsDuplicate = reported.IsIdempotentReplay,
});
}
private static bool TryNormalizeReport(
ReportConnectionOutcomeRequest request,
out ConnectionOutcomeKind outcome,
out ConnectionElapsedBucket elapsedBucket)
{
outcome = request.Outcome switch
{
ConnectionOutcomeKind.TimedOut => ConnectionOutcomeKind.PunchTimedOut,
ConnectionOutcomeKind.StaleHost => ConnectionOutcomeKind.NoHostPresence,
ConnectionOutcomeKind.TransportFailed => ConnectionOutcomeKind.TransportError,
_ => request.Outcome,
};
if (!ContractValidation.IsReportableConnectionOutcome(request.Outcome))
{
elapsedBucket = default;
return false;
}
if (Enum.IsDefined(request.ElapsedBucket))
{
elapsedBucket = request.ElapsedBucket;
return true;
}
#pragma warning disable CS0618 // Frozen v1 compatibility input; never retained at exact precision.
if (request.ElapsedBucket == default && request.ElapsedMilliseconds >= 0)
{
elapsedBucket = BucketElapsedMilliseconds(request.ElapsedMilliseconds);
return true;
}
#pragma warning restore CS0618
elapsedBucket = default;
return false;
}
private static ConnectionElapsedBucket BucketElapsedMilliseconds(int elapsedMilliseconds) =>
elapsedMilliseconds switch
{
< 1_000 => ConnectionElapsedBucket.UnderOneSecond,
< 5_000 => ConnectionElapsedBucket.OneToFiveSeconds,
< 15_000 => ConnectionElapsedBucket.FiveToFifteenSeconds,
< 30_000 => ConnectionElapsedBucket.FifteenToThirtySeconds,
_ => ConnectionElapsedBucket.ThirtySecondsOrMore,
};
}
@@ -0,0 +1,235 @@
using System.ComponentModel.DataAnnotations;
using System.Net;
using FinalFactory.Rendezvous.Server.Abuse;
namespace FinalFactory.Rendezvous.Server.Deployment;
internal sealed record DeploymentOptions
{
public const string SectionName = "Rendezvous:Deployment";
[Required]
public string PublicHttpBaseUrl { get; init; } = string.Empty;
[Required]
public string PublicUdpHost { get; init; } = string.Empty;
[Range(1, 65_535)]
public int PublicUdpPort { get; init; } = 9050;
[Range(1, 30)]
public int DrainDeadlineSeconds { get; init; } = 30;
[Range(0, 5)]
public int MinimumDrainSeconds { get; init; } = 1;
public bool SingleActiveInstance { get; init; } = true;
public bool AllowPrivatePublicEndpoints { get; init; }
public IReadOnlyList<string> ValidateProduction(
AbuseProtectionOptions abuseProtection,
string? allowedHosts)
{
List<string> errors = [];
if (!SingleActiveInstance)
{
errors.Add("Rendezvous:Deployment:SingleActiveInstance must be true because ephemeral state is not shared between replicas.");
}
if (MinimumDrainSeconds >= DrainDeadlineSeconds)
{
errors.Add("Rendezvous:Deployment:MinimumDrainSeconds must be less than DrainDeadlineSeconds.");
}
if (DrainDeadlineSeconds is < 1 or > 30)
{
errors.Add("Rendezvous:Deployment:DrainDeadlineSeconds must be between 1 and 30.");
}
if (MinimumDrainSeconds is < 0 or > 5)
{
errors.Add("Rendezvous:Deployment:MinimumDrainSeconds must be between 0 and 5.");
}
if (PublicUdpPort is < 1 or > 65_535)
{
errors.Add("Rendezvous:Deployment:PublicUdpPort must be between 1 and 65535.");
}
ValidateHttpEndpoint(errors);
ValidateUdpEndpoint(errors);
ValidateAllowedHosts(errors, allowedHosts, PublicHttpBaseUrl);
if (abuseProtection.TrustedProxyAddresses is not { Length: > 0 })
{
errors.Add(
"Rendezvous:AbuseProtection:TrustedProxyAddresses must list the exact TLS proxy addresses; forwarded headers are rejected without this trust boundary.");
}
return errors;
}
private void ValidateHttpEndpoint(List<string> errors)
{
if (!Uri.TryCreate(PublicHttpBaseUrl, UriKind.Absolute, out Uri? endpoint)
|| !string.Equals(endpoint.Scheme, Uri.UriSchemeHttps, StringComparison.Ordinal)
|| !string.IsNullOrEmpty(endpoint.UserInfo)
|| !string.IsNullOrEmpty(endpoint.Query)
|| !string.IsNullOrEmpty(endpoint.Fragment)
|| endpoint.AbsolutePath != "/")
{
errors.Add(
"Rendezvous:Deployment:PublicHttpBaseUrl must be an absolute HTTPS origin with no credentials, path, query, or fragment (for example, https://rendezvous.your-company.tld/).");
return;
}
if (!AllowPrivatePublicEndpoints && !IsPublicHost(endpoint.Host))
{
errors.Add(
"Rendezvous:Deployment:PublicHttpBaseUrl must use a public DNS name or address; set AllowPrivatePublicEndpoints=true only for an isolated deployment smoke test.");
}
}
private void ValidateUdpEndpoint(List<string> errors)
{
if (string.IsNullOrWhiteSpace(PublicUdpHost)
|| PublicUdpHost.Contains("//", StringComparison.Ordinal)
|| PublicUdpHost.Contains(':', StringComparison.Ordinal) && !IPAddress.TryParse(PublicUdpHost, out _)
|| Uri.CheckHostName(PublicUdpHost) == UriHostNameType.Unknown)
{
errors.Add(
"Rendezvous:Deployment:PublicUdpHost must contain only the advertised DNS name or IP address; configure the port separately.");
return;
}
if (!AllowPrivatePublicEndpoints && !IsPublicHost(PublicUdpHost))
{
errors.Add(
"Rendezvous:Deployment:PublicUdpHost must use a public DNS name or address; set AllowPrivatePublicEndpoints=true only for an isolated deployment smoke test.");
}
}
private static void ValidateAllowedHosts(
List<string> errors,
string? allowedHosts,
string publicHttpBaseUrl)
{
string[] hosts = (allowedHosts ?? string.Empty).Split(
';',
StringSplitOptions.RemoveEmptyEntries | StringSplitOptions.TrimEntries);
if (hosts.Length == 0 || hosts.Any(static host => host is "*" or "+"))
{
errors.Add(
"AllowedHosts must explicitly list the public HTTP host in production; wildcard or empty host filtering is unsafe.");
return;
}
if (Uri.TryCreate(publicHttpBaseUrl, UriKind.Absolute, out Uri? endpoint)
&& !hosts.Contains(endpoint.Host, StringComparer.OrdinalIgnoreCase))
{
errors.Add(
"AllowedHosts must contain the exact host advertised by Rendezvous:Deployment:PublicHttpBaseUrl.");
}
}
private static bool IsPublicHost(string host)
{
if (!IPAddress.TryParse(host, out IPAddress? address))
{
return Uri.CheckHostName(host) == UriHostNameType.Dns
&& host.Contains('.', StringComparison.Ordinal)
&& !IsReservedDnsName(host);
}
return IsGloballyRoutableUnicast(address);
}
private static bool IsReservedDnsName(string host)
{
string normalized = host.TrimEnd('.');
string[] reservedSuffixes =
[
"localhost",
"local",
"invalid",
"test",
"example",
"example.com",
"example.net",
"example.org",
"home.arpa",
"alt",
"onion",
];
return reservedSuffixes.Any(suffix =>
string.Equals(normalized, suffix, StringComparison.OrdinalIgnoreCase)
|| normalized.EndsWith($".{suffix}", StringComparison.OrdinalIgnoreCase));
}
private static bool IsGloballyRoutableUnicast(IPAddress address)
{
if (address.IsIPv4MappedToIPv6)
{
address = address.MapToIPv4();
}
byte[] bytes = address.GetAddressBytes();
if (address.AddressFamily == System.Net.Sockets.AddressFamily.InterNetwork)
{
return !(bytes[0] is 0 or 10 or 127
|| bytes[0] == 100 && bytes[1] is >= 64 and <= 127
|| bytes[0] == 169 && bytes[1] == 254
|| bytes[0] == 172 && bytes[1] is >= 16 and <= 31
|| bytes[0] == 192
&& (bytes[1] == 0 && bytes[2] is 0 or 2
|| bytes[1] == 88 && bytes[2] == 99
|| bytes[1] == 168)
|| bytes[0] == 198
&& (bytes[1] is 18 or 19
|| bytes[1] == 51 && bytes[2] == 100)
|| bytes[0] == 203 && bytes[1] == 0 && bytes[2] == 113
|| bytes[0] >= 224);
}
return address.AddressFamily == System.Net.Sockets.AddressFamily.InterNetworkV6
&& !IPAddress.IsLoopback(address)
&& !address.Equals(IPAddress.IPv6Any)
&& !address.IsIPv6LinkLocal
&& !address.IsIPv6SiteLocal
&& !address.IsIPv6Multicast
&& (bytes[0] & 0xe0) == 0x20
&& !HasPrefix(bytes, [0x20, 0x01, 0x00], 23)
&& !HasPrefix(bytes, [0x20, 0x01, 0x0d, 0xb8], 32)
&& !HasPrefix(bytes, [0x20, 0x02], 16)
&& !HasPrefix(bytes, [0x3f, 0xff, 0x00], 20);
}
private static bool HasPrefix(byte[] address, byte[] prefix, int bitCount)
{
int fullBytes = bitCount / 8;
for (int index = 0; index < fullBytes; index++)
{
if (address[index] != prefix[index])
{
return false;
}
}
int remainingBits = bitCount % 8;
if (remainingBits == 0)
{
return true;
}
int mask = 0xff << (8 - remainingBits);
return (address[fullBytes] & mask) == (prefix[fullBytes] & mask);
}
}
internal sealed class DeploymentConfigurationException(IReadOnlyList<string> errors)
: InvalidOperationException(
"Production deployment configuration is invalid:" + Environment.NewLine
+ string.Join(Environment.NewLine, errors.Select(static error => $"- {error}")))
{
}
@@ -0,0 +1,102 @@
using System.Diagnostics;
using FinalFactory.Rendezvous.Server.State;
using Microsoft.Extensions.Options;
namespace FinalFactory.Rendezvous.Server.Deployment;
internal sealed class GracefulDrainService : IHostedService, IDisposable
{
private static readonly TimeSpan PollInterval = TimeSpan.FromMilliseconds(50);
private readonly InMemoryEphemeralRendezvousStore _store;
private readonly IHostApplicationLifetime _lifetime;
private readonly DeploymentOptions _options;
private readonly ILogger<GracefulDrainService> _logger;
private readonly object _gate = new();
private CancellationTokenRegistration _stoppingRegistration;
private Task? _drainTask;
public GracefulDrainService(
InMemoryEphemeralRendezvousStore store,
IHostApplicationLifetime lifetime,
IOptions<DeploymentOptions> options,
ILogger<GracefulDrainService> logger)
{
_store = store;
_lifetime = lifetime;
_options = options.Value;
_logger = logger;
}
public Task StartAsync(CancellationToken cancellationToken)
{
cancellationToken.ThrowIfCancellationRequested();
_stoppingRegistration = _lifetime.ApplicationStopping.Register(
() => EnsureDrainAsync().GetAwaiter().GetResult());
return Task.CompletedTask;
}
public Task StopAsync(CancellationToken cancellationToken)
{
// ApplicationStopping callbacks run before hosted services and listeners
// stop. StopAsync is the idempotent fallback for directly driven hosts.
_ = cancellationToken;
return EnsureDrainAsync();
}
public void Dispose() => _stoppingRegistration.Dispose();
private Task EnsureDrainAsync()
{
lock (_gate)
{
return _drainTask ??= DrainAsync();
}
}
private async Task DrainAsync()
{
_store.BeginDrain(CancellationToken.None);
TimeSpan deadline = TimeSpan.FromSeconds(_options.DrainDeadlineSeconds);
TimeSpan minimum = TimeSpan.FromSeconds(_options.MinimumDrainSeconds);
long startedAt = Stopwatch.GetTimestamp();
LogDrainStarted(_logger, _options.DrainDeadlineSeconds);
try
{
while (Stopwatch.GetElapsedTime(startedAt) < deadline)
{
TimeSpan elapsed = Stopwatch.GetElapsedTime(startedAt);
if (elapsed >= minimum && _store.GetActiveJoinAttemptCountForDrain() == 0)
{
break;
}
TimeSpan remaining = deadline - elapsed;
await Task.Delay(
remaining < PollInterval ? remaining : PollInterval,
CancellationToken.None).ConfigureAwait(false);
}
}
finally
{
_store.MarkUnavailable();
double elapsedMilliseconds = Stopwatch.GetElapsedTime(startedAt).TotalMilliseconds;
LogDrainFinished(_logger, elapsedMilliseconds);
}
}
private static readonly Action<ILogger, int, Exception?> DrainStarted = LoggerMessage.Define<int>(
LogLevel.Information,
new EventId(1, nameof(LogDrainStarted)),
"Graceful drain started with a {DrainDeadlineSeconds}-second deadline");
private static readonly Action<ILogger, double, Exception?> DrainFinished = LoggerMessage.Define<double>(
LogLevel.Information,
new EventId(2, nameof(LogDrainFinished)),
"Graceful drain finished after {ElapsedMilliseconds:F0} ms; ephemeral state was cleared");
private static void LogDrainStarted(ILogger logger, int drainDeadlineSeconds) =>
DrainStarted(logger, drainDeadlineSeconds, null);
private static void LogDrainFinished(ILogger logger, double elapsedMilliseconds) =>
DrainFinished(logger, elapsedMilliseconds, null);
}
@@ -3,6 +3,7 @@
<TargetFramework>net10.0</TargetFramework> <TargetFramework>net10.0</TargetFramework>
<AssemblyName>FinalFactory.Rendezvous.Server</AssemblyName> <AssemblyName>FinalFactory.Rendezvous.Server</AssemblyName>
<RootNamespace>FinalFactory.Rendezvous.Server</RootNamespace> <RootNamespace>FinalFactory.Rendezvous.Server</RootNamespace>
<EnableDefaultCompileItems>false</EnableDefaultCompileItems>
<IsPackable>false</IsPackable> <IsPackable>false</IsPackable>
<OpenApiGenerateDocuments>true</OpenApiGenerateDocuments> <OpenApiGenerateDocuments>true</OpenApiGenerateDocuments>
<OpenApiDocumentsDirectory>$(MSBuildProjectDirectory)/../../docs/api</OpenApiDocumentsDirectory> <OpenApiDocumentsDirectory>$(MSBuildProjectDirectory)/../../docs/api</OpenApiDocumentsDirectory>
@@ -14,4 +15,74 @@
<PackageReference Include="Microsoft.AspNetCore.OpenApi" /> <PackageReference Include="Microsoft.AspNetCore.OpenApi" />
<PackageReference Include="Microsoft.Extensions.ApiDescription.Server" PrivateAssets="all" /> <PackageReference Include="Microsoft.Extensions.ApiDescription.Server" PrivateAssets="all" />
</ItemGroup> </ItemGroup>
<ItemGroup>
<AssemblyAttribute Include="System.Reflection.AssemblyMetadataAttribute">
<_Parameter1>RendezvousMinimumClientVersion</_Parameter1>
<_Parameter2>$(MinimumClientVersion)</_Parameter2>
</AssemblyAttribute>
<AssemblyAttribute Include="System.Reflection.AssemblyMetadataAttribute">
<_Parameter1>RendezvousMaximumClientMajorVersion</_Parameter1>
<_Parameter2>$(MaximumClientMajorVersion)</_Parameter2>
</AssemblyAttribute>
<AssemblyAttribute Include="System.Reflection.AssemblyMetadataAttribute">
<_Parameter1>RendezvousUdpContractVersion</_Parameter1>
<_Parameter2>$(UdpContractVersion)</_Parameter2>
</AssemblyAttribute>
<AssemblyAttribute Include="System.Reflection.AssemblyMetadataAttribute">
<_Parameter1>RendezvousConnectionTicketFormatVersion</_Parameter1>
<_Parameter2>$(ConnectionTicketFormatVersion)</_Parameter2>
</AssemblyAttribute>
<AssemblyAttribute Include="System.Reflection.AssemblyMetadataAttribute">
<_Parameter1>RendezvousLiteNetLibMajorVersion</_Parameter1>
<_Parameter2>$(LiteNetLibMajorVersion)</_Parameter2>
</AssemblyAttribute>
</ItemGroup>
<ItemGroup>
<!-- Roslyn and source generators consume syntax trees in item order. Keep
this ordinal manifest explicit so clean builds are byte reproducible. -->
<Compile Include="Abuse/AbuseProtectionOptions.cs" />
<Compile Include="Abuse/AbuseProtectionService.cs" />
<Compile Include="Abuse/HttpAbuseProtectionMiddleware.cs" />
<Compile Include="Abuse/TrustedProxyForwarding.cs" />
<Compile Include="Browser/EphemeralCursorProtector.cs" />
<Compile Include="Browser/SessionBrowserCursorCodec.cs" />
<Compile Include="Browser/SessionBrowserService.cs" />
<Compile Include="ConnectionOutcomes/ConnectionOutcomeService.cs" />
<Compile Include="Deployment/DeploymentOptions.cs" />
<Compile Include="Deployment/GracefulDrainService.cs" />
<Compile Include="Http/ContractEndpoints.cs" />
<Compile Include="Http/RendezvousExceptionHandler.cs" />
<Compile Include="JoinAttempts/JoinAttemptCursorCodec.cs" />
<Compile Include="JoinAttempts/JoinAttemptService.cs" />
<Compile Include="Observability/AuditOptions.cs" />
<Compile Include="Observability/AuditTrail.cs" />
<Compile Include="Observability/HealthEndpoints.cs" />
<Compile Include="Observability/RendezvousReadiness.cs" />
<Compile Include="Observability/RendezvousTelemetry.cs" />
<Compile Include="Observability/TelemetryMiddleware.cs" />
<Compile Include="Operations/OperatorEndpoints.cs" />
<Compile Include="Operations/OperatorModels.cs" />
<Compile Include="Operations/OperatorService.cs" />
<Compile Include="Operations/ReleaseCompatibility.cs" />
<Compile Include="Program.cs" />
<Compile Include="Properties/AssemblyInfo.cs" />
<Compile Include="Provisioning/GamePolicy.cs" />
<Compile Include="Provisioning/GamePolicyRegistry.cs" />
<Compile Include="Provisioning/PrincipalCredentialService.cs" />
<Compile Include="Provisioning/Principals.cs" />
<Compile Include="Provisioning/ProvisioningOptions.cs" />
<Compile Include="Provisioning/ProvisioningRuntime.cs" />
<Compile Include="Provisioning/PublisherAuthorizationService.cs" />
<Compile Include="Provisioning/SecretProviders.cs" />
<Compile Include="Provisioning/SigningKeyRing.cs" />
<Compile Include="Sessions/EphemeralCapabilityIssuer.cs" />
<Compile Include="Sessions/SessionLeaseService.cs" />
<Compile Include="State/EphemeralStateContracts.cs" />
<Compile Include="State/InMemoryEphemeralRendezvousStore.cs" />
<Compile Include="State/StoreResultMapping.cs" />
<Compile Include="Transport/LiteNetNatRequestCodec.cs" />
<Compile Include="Transport/NatMediationProcessor.cs" />
<Compile Include="Transport/UdpMediatorOptions.cs" />
<Compile Include="Transport/UdpMediatorService.cs" />
</ItemGroup>
</Project> </Project>
@@ -1,5 +1,9 @@
using System.Net;
using FinalFactory.Rendezvous.Contracts; using FinalFactory.Rendezvous.Contracts;
using FinalFactory.Rendezvous.Server.Abuse;
using FinalFactory.Rendezvous.Server.Browser; using FinalFactory.Rendezvous.Server.Browser;
using FinalFactory.Rendezvous.Server.ConnectionOutcomes;
using FinalFactory.Rendezvous.Server.JoinAttempts;
using FinalFactory.Rendezvous.Server.Provisioning; using FinalFactory.Rendezvous.Server.Provisioning;
using FinalFactory.Rendezvous.Server.Sessions; using FinalFactory.Rendezvous.Server.Sessions;
using FinalFactory.Rendezvous.Server.State; using FinalFactory.Rendezvous.Server.State;
@@ -9,8 +13,6 @@ namespace FinalFactory.Rendezvous.Server.Http;
internal static class ContractEndpoints internal static class ContractEndpoints
{ {
private const int NotImplementedStatus = StatusCodes.Status501NotImplemented;
public static IEndpointRouteBuilder MapRendezvousContractEndpoints( public static IEndpointRouteBuilder MapRendezvousContractEndpoints(
this IEndpointRouteBuilder endpoints) this IEndpointRouteBuilder endpoints)
{ {
@@ -19,6 +21,7 @@ internal static class ContractEndpoints
.Accepts<RegisterSessionRequest>("application/json") .Accepts<RegisterSessionRequest>("application/json")
.Produces<RegisterSessionResponse>(StatusCodes.Status201Created) .Produces<RegisterSessionResponse>(StatusCodes.Status201Created)
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized) .Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden) .Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status409Conflict) .Produces<ApiError>(StatusCodes.Status409Conflict)
@@ -30,44 +33,55 @@ internal static class ContractEndpoints
.Accepts<RenewLeaseRequest>("application/json") .Accepts<RenewLeaseRequest>("application/json")
.Produces<RenewLeaseResponse>() .Produces<RenewLeaseResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized) .Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden) .Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound) .Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status409Conflict) .Produces<ApiError>(StatusCodes.Status409Conflict)
.Produces<ApiError>(StatusCodes.Status410Gone) .Produces<ApiError>(StatusCodes.Status410Gone)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable) .Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("RenewSessionLease"); .WithName("RenewSessionLease");
sessions.MapPut("/{listingId}", UpdateSession) sessions.MapPut("/{listingId}", UpdateSession)
.Accepts<UpdateSessionRequest>("application/json") .Accepts<UpdateSessionRequest>("application/json")
.Produces(StatusCodes.Status204NoContent) .Produces(StatusCodes.Status204NoContent)
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized) .Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden) .Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound) .Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable) .Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("UpdateSession"); .WithName("UpdateSession");
sessions.MapDelete("/{listingId}", DeleteSession) sessions.MapDelete("/{listingId}", DeleteSession)
.Accepts<DeleteSessionRequest>("application/json") .Accepts<DeleteSessionRequest>("application/json")
.Produces(StatusCodes.Status204NoContent) .Produces(StatusCodes.Status204NoContent)
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized) .Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden) .Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable) .Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("DeleteSession"); .WithName("DeleteSession");
sessions.MapGet("/", BrowseSessions) sessions.MapGet("/", BrowseSessions)
.Produces<BrowseSessionsResponse>() .Produces<BrowseSessionsResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable) .Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("BrowseSessions"); .WithName("BrowseSessions");
sessions.MapGet("/{listingId}", GetSession) sessions.MapGet("/{listingId}", GetSession)
.Produces<GetSessionResponse>() .Produces<GetSessionResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status404NotFound) .Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable) .Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("GetSession"); .WithName("GetSession");
sessions.MapGet("/{listingId}/join-attempts", BrowseHostJoinAttempts) sessions.MapGet("/{listingId}/join-attempts", BrowseHostJoinAttempts)
.Produces<BrowseHostJoinAttemptsResponse>() .Produces<BrowseHostJoinAttemptsResponse>()
.Produces<ApiError>(NotImplementedStatus) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("BrowseHostJoinAttempts"); .WithName("BrowseHostJoinAttempts");
RouteGroupBuilder attempts = endpoints RouteGroupBuilder attempts = endpoints
@@ -76,12 +90,30 @@ internal static class ContractEndpoints
attempts.MapPost("/", CreateJoinAttempt) attempts.MapPost("/", CreateJoinAttempt)
.Accepts<CreateJoinAttemptRequest>("application/json") .Accepts<CreateJoinAttemptRequest>("application/json")
.Produces<CreateJoinAttemptResponse>(StatusCodes.Status201Created) .Produces<CreateJoinAttemptResponse>(StatusCodes.Status201Created)
.Produces<ApiError>(NotImplementedStatus) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status409Conflict)
.Produces<ApiError>(StatusCodes.Status410Gone)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("CreateJoinAttempt"); .WithName("CreateJoinAttempt");
attempts.MapDelete("/{attemptId}", CancelJoinAttempt)
.Produces(StatusCodes.Status204NoContent)
.Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("CancelJoinAttempt");
attempts.MapPost("/{attemptId}/outcome", ReportConnectionOutcome) attempts.MapPost("/{attemptId}/outcome", ReportConnectionOutcome)
.Accepts<ReportConnectionOutcomeRequest>("application/json") .Accepts<ReportConnectionOutcomeRequest>("application/json")
.Produces<ReportConnectionOutcomeResponse>() .Produces<ReportConnectionOutcomeResponse>()
.Produces<ApiError>(NotImplementedStatus) .Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status409Conflict)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("ReportConnectionOutcome"); .WithName("ReportConnectionOutcome");
return endpoints; return endpoints;
@@ -92,6 +124,7 @@ internal static class ContractEndpoints
[FromHeader(Name = "Authorization")] string? authorizationHeader, [FromHeader(Name = "Authorization")] string? authorizationHeader,
[FromServices] PrincipalCredentialService credentials, [FromServices] PrincipalCredentialService credentials,
[FromServices] SessionLeaseService sessions, [FromServices] SessionLeaseService sessions,
[FromServices] AbuseProtectionService abuseProtection,
[FromServices] IWallClock clock, [FromServices] IWallClock clock,
HttpContext httpContext, HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
@@ -105,6 +138,21 @@ internal static class ContractEndpoints
return AuthenticationRequired(httpContext); return AuthenticationRequired(httpContext);
} }
IPublisherPrincipal publisher = (IPublisherPrincipal)principal!;
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"RegisterSession",
Tenant(publisher.GameId, publisher.EnvironmentId),
publisher.Subject,
null,
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
SessionServiceResult<RegisterSessionResponse> result = sessions.Register( SessionServiceResult<RegisterSessionResponse> result = sessions.Register(
principal!, principal!,
request, request,
@@ -113,6 +161,7 @@ internal static class ContractEndpoints
? Results.Created($"/v1/sessions/{result.Value.ListingId}", result.Value) ? Results.Created($"/v1/sessions/{result.Value.ListingId}", result.Value)
: Error(result.Error); : Error(result.Error);
} }
}
private static IResult RenewLease( private static IResult RenewLease(
SessionListingId listingId, SessionListingId listingId,
@@ -120,6 +169,7 @@ internal static class ContractEndpoints
[FromHeader(Name = "Authorization")] string? authorizationHeader, [FromHeader(Name = "Authorization")] string? authorizationHeader,
[FromServices] PrincipalCredentialService credentials, [FromServices] PrincipalCredentialService credentials,
[FromServices] SessionLeaseService sessions, [FromServices] SessionLeaseService sessions,
[FromServices] AbuseProtectionService abuseProtection,
[FromServices] IWallClock clock, [FromServices] IWallClock clock,
HttpContext httpContext, HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
@@ -133,6 +183,21 @@ internal static class ContractEndpoints
return AuthenticationRequired(httpContext); return AuthenticationRequired(httpContext);
} }
IPublisherPrincipal publisher = (IPublisherPrincipal)principal!;
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"RenewSessionLease",
Tenant(publisher.GameId, publisher.EnvironmentId),
publisher.Subject,
listingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
SessionServiceResult<RenewLeaseResponse> result = sessions.Renew( SessionServiceResult<RenewLeaseResponse> result = sessions.Renew(
principal!, principal!,
listingId, listingId,
@@ -142,6 +207,7 @@ internal static class ContractEndpoints
? Results.Ok(result.Value) ? Results.Ok(result.Value)
: Error(result.Error); : Error(result.Error);
} }
}
private static IResult UpdateSession( private static IResult UpdateSession(
SessionListingId listingId, SessionListingId listingId,
@@ -149,6 +215,7 @@ internal static class ContractEndpoints
[FromHeader(Name = "Authorization")] string? authorizationHeader, [FromHeader(Name = "Authorization")] string? authorizationHeader,
[FromServices] PrincipalCredentialService credentials, [FromServices] PrincipalCredentialService credentials,
[FromServices] SessionLeaseService sessions, [FromServices] SessionLeaseService sessions,
[FromServices] AbuseProtectionService abuseProtection,
[FromServices] IWallClock clock, [FromServices] IWallClock clock,
HttpContext httpContext, HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
@@ -162,6 +229,21 @@ internal static class ContractEndpoints
return AuthenticationRequired(httpContext); return AuthenticationRequired(httpContext);
} }
IPublisherPrincipal publisher = (IPublisherPrincipal)principal!;
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"UpdateSession",
Tenant(publisher.GameId, publisher.EnvironmentId),
publisher.Subject,
listingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
SessionServiceResult<bool> result = sessions.Update( SessionServiceResult<bool> result = sessions.Update(
principal!, principal!,
listingId, listingId,
@@ -169,6 +251,7 @@ internal static class ContractEndpoints
cancellationToken); cancellationToken);
return result.Succeeded ? Results.NoContent() : Error(result.Error); return result.Succeeded ? Results.NoContent() : Error(result.Error);
} }
}
private static IResult DeleteSession( private static IResult DeleteSession(
SessionListingId listingId, SessionListingId listingId,
@@ -176,6 +259,7 @@ internal static class ContractEndpoints
[FromHeader(Name = "Authorization")] string? authorizationHeader, [FromHeader(Name = "Authorization")] string? authorizationHeader,
[FromServices] PrincipalCredentialService credentials, [FromServices] PrincipalCredentialService credentials,
[FromServices] SessionLeaseService sessions, [FromServices] SessionLeaseService sessions,
[FromServices] AbuseProtectionService abuseProtection,
[FromServices] IWallClock clock, [FromServices] IWallClock clock,
HttpContext httpContext, HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
@@ -189,6 +273,21 @@ internal static class ContractEndpoints
return AuthenticationRequired(httpContext); return AuthenticationRequired(httpContext);
} }
IPublisherPrincipal publisher = (IPublisherPrincipal)principal!;
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"DeleteSession",
Tenant(publisher.GameId, publisher.EnvironmentId),
publisher.Subject,
listingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
SessionServiceResult<bool> result = sessions.Delete( SessionServiceResult<bool> result = sessions.Delete(
principal!, principal!,
listingId, listingId,
@@ -196,6 +295,7 @@ internal static class ContractEndpoints
cancellationToken); cancellationToken);
return result.Succeeded ? Results.NoContent() : Error(result.Error); return result.Succeeded ? Results.NoContent() : Error(result.Error);
} }
}
private static IResult BrowseSessions( private static IResult BrowseSessions(
[FromQuery] int contractVersion, [FromQuery] int contractVersion,
@@ -207,6 +307,8 @@ internal static class ContractEndpoints
[FromQuery] bool? excludeFull, [FromQuery] bool? excludeFull,
[FromQuery] string? cursor, [FromQuery] string? cursor,
[FromServices] SessionBrowserService browser, [FromServices] SessionBrowserService browser,
[FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
{ {
if (!GameId.TryParse(gameId, out GameId parsedGameId) if (!GameId.TryParse(gameId, out GameId parsedGameId)
@@ -216,6 +318,20 @@ internal static class ContractEndpoints
return Error(RendezvousErrorCode.InvalidRequest); return Error(RendezvousErrorCode.InvalidRequest);
} }
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"BrowseSessions",
Tenant(parsedGameId, parsedEnvironmentId),
null,
null,
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
BrowserServiceResult<BrowseSessionsResponse> result = browser.Browse(new() BrowserServiceResult<BrowseSessionsResponse> result = browser.Browse(new()
{ {
ContractVersion = contractVersion, ContractVersion = contractVersion,
@@ -231,6 +347,7 @@ internal static class ContractEndpoints
? Results.Ok(result.Value) ? Results.Ok(result.Value)
: Error(result.Error); : Error(result.Error);
} }
}
private static IResult GetSession( private static IResult GetSession(
SessionListingId listingId, SessionListingId listingId,
@@ -239,6 +356,8 @@ internal static class ContractEndpoints
[FromQuery] string environmentId, [FromQuery] string environmentId,
[FromQuery] uint protocolVersion, [FromQuery] uint protocolVersion,
[FromServices] SessionBrowserService browser, [FromServices] SessionBrowserService browser,
[FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken) CancellationToken cancellationToken)
{ {
if (ContractValidation.ValidateContractVersion(contractVersion) != RendezvousErrorCode.None) if (ContractValidation.ValidateContractVersion(contractVersion) != RendezvousErrorCode.None)
@@ -252,6 +371,20 @@ internal static class ContractEndpoints
return Error(RendezvousErrorCode.InvalidRequest); return Error(RendezvousErrorCode.InvalidRequest);
} }
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"GetSession",
Tenant(parsedGameId, parsedEnvironmentId),
null,
listingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
BrowserServiceResult<GetSessionResponse> result = browser.Get( BrowserServiceResult<GetSessionResponse> result = browser.Get(
listingId, listingId,
parsedGameId, parsedGameId,
@@ -262,29 +395,146 @@ internal static class ContractEndpoints
? Results.Ok(result.Value) ? Results.Ok(result.Value)
: Error(result.Error); : Error(result.Error);
} }
}
private static IResult BrowseHostJoinAttempts( private static IResult BrowseHostJoinAttempts(
SessionListingId listingId, SessionListingId listingId,
[FromQuery] int contractVersion, [FromQuery] int contractVersion,
[FromHeader(Name = "X-Rendezvous-Lease-Token")] string leaseToken, [FromHeader(Name = "X-Rendezvous-Lease-Token")] string leaseToken,
[FromQuery] int? pageSize, [FromQuery] int? pageSize,
[FromQuery] string? cursor) => NotImplemented(); [FromQuery] string? cursor,
[FromServices] JoinAttemptService attempts,
[FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken)
{
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"BrowseHostJoinAttempts",
null,
AbuseProtectionService.FingerprintSecret(leaseToken ?? string.Empty),
listingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
private static IResult CreateJoinAttempt([FromBody] CreateJoinAttemptRequest request) => using (abuseLease)
NotImplemented(); {
JoinAttemptServiceResult<BrowseHostJoinAttemptsResponse> result = attempts.BrowseForHost(
listingId,
contractVersion,
leaseToken,
pageSize ?? ContractLimits.BrowserPageMaxItems,
cursor,
cancellationToken);
return result.Succeeded && result.Value is not null
? Results.Ok(result.Value)
: Error(result.Error);
}
}
private static IResult CreateJoinAttempt(
[FromBody] CreateJoinAttemptRequest request,
[FromServices] JoinAttemptService attempts,
[FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken)
{
if (httpContext.Connection.RemoteIpAddress is not IPAddress remoteAddress)
{
return Error(RendezvousErrorCode.InvalidRequest);
}
string clientSubject = attempts.CreateAnonymousClientSubject(remoteAddress);
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"CreateJoinAttempt",
Tenant(request.GameId, request.EnvironmentId),
clientSubject,
request.ListingId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
JoinAttemptServiceResult<CreateJoinAttemptResponse> result = attempts.Create(
clientSubject,
request,
cancellationToken);
return result.Succeeded && result.Value is not null
? Results.Created($"/v1/join-attempts/{result.Value.AttemptId}", result.Value)
: Error(result.Error);
}
}
private static IResult CancelJoinAttempt(
JoinAttemptId attemptId,
[FromHeader(Name = "X-Rendezvous-Client-Punch-Capability")] string clientPunchCapability,
[FromServices] JoinAttemptService attempts,
[FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken)
{
if (!TryAcquireIdentity(
abuseProtection,
httpContext,
"CancelJoinAttempt",
null,
AbuseProtectionService.FingerprintSecret(clientPunchCapability ?? string.Empty),
attemptId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
JoinAttemptServiceResult<bool> result = attempts.Cancel(
attemptId,
clientPunchCapability,
cancellationToken);
return result.Succeeded ? Results.NoContent() : Error(result.Error);
}
}
private static IResult ReportConnectionOutcome( private static IResult ReportConnectionOutcome(
JoinAttemptId attemptId, JoinAttemptId attemptId,
[FromBody] ReportConnectionOutcomeRequest request) => NotImplemented(); [FromHeader(Name = "X-Rendezvous-Client-Punch-Capability")] string clientPunchCapability,
[FromBody] ReportConnectionOutcomeRequest request,
private static IResult NotImplemented() => Results.Json( [FromServices] ConnectionOutcomeService outcomes,
new ApiError [FromServices] AbuseProtectionService abuseProtection,
HttpContext httpContext,
CancellationToken cancellationToken)
{ {
Code = RendezvousErrorCode.ServiceUnavailable, if (!TryAcquireIdentity(
Message = "The v1 contract is reserved; implementation is tracked by subsequent issues.", abuseProtection,
}, httpContext,
ContractJson.Options, "ReportConnectionOutcome",
statusCode: NotImplementedStatus); null,
AbuseProtectionService.FingerprintSecret(clientPunchCapability ?? string.Empty),
attemptId.ToString(),
out AbuseProtectionService.AbuseLease? abuseLease))
{
return RateLimited(httpContext);
}
using (abuseLease)
{
ConnectionOutcomeServiceResult result = outcomes.Report(
attemptId,
clientPunchCapability,
request,
cancellationToken);
return result.Succeeded && result.Value is not null
? Results.Ok(result.Value)
: Error(result.Error);
}
}
private static bool TryAuthenticatePublisher( private static bool TryAuthenticatePublisher(
string? authorizationHeader, string? authorizationHeader,
@@ -311,11 +561,41 @@ internal static class ContractEndpoints
return true; return true;
} }
private static IResult Error(RendezvousErrorCode code) => Results.Json( private static bool TryAcquireIdentity(
AbuseProtectionService abuseProtection,
HttpContext httpContext,
string operation,
string? tenant,
string? principal,
string? resource,
out AbuseProtectionService.AbuseLease? lease)
{
if (abuseProtection.TryAcquireHttpIdentity(
operation,
httpContext.Connection.RemoteIpAddress,
tenant,
principal,
resource,
out lease,
out int retryAfterSeconds))
{
return true;
}
httpContext.Response.Headers.RetryAfter = retryAfterSeconds.ToString(
System.Globalization.CultureInfo.InvariantCulture);
return false;
}
private static string Tenant(GameId gameId, EnvironmentId environmentId) =>
$"{gameId.Value}/{environmentId.Value}";
private static IResult Error(RendezvousErrorCode code, int? retryAfterSeconds = null) => Results.Json(
new ApiError new ApiError
{ {
Code = code, Code = code,
Message = ErrorMessage(code), Message = ErrorMessage(code),
RetryAfterSeconds = retryAfterSeconds,
}, },
ContractJson.Options, ContractJson.Options,
statusCode: ErrorStatus(code)); statusCode: ErrorStatus(code));
@@ -326,13 +606,27 @@ internal static class ContractEndpoints
return Error(RendezvousErrorCode.AuthenticationRequired); return Error(RendezvousErrorCode.AuthenticationRequired);
} }
private static IResult RateLimited(HttpContext context)
{
int? retryAfterSeconds = int.TryParse(
context.Response.Headers.RetryAfter,
System.Globalization.NumberStyles.None,
System.Globalization.CultureInfo.InvariantCulture,
out int parsed)
? Math.Clamp(parsed, 1, 60)
: null;
return Error(RendezvousErrorCode.RateLimited, retryAfterSeconds);
}
private static int ErrorStatus(RendezvousErrorCode code) => code switch private static int ErrorStatus(RendezvousErrorCode code) => code switch
{ {
RendezvousErrorCode.AuthenticationRequired => StatusCodes.Status401Unauthorized, RendezvousErrorCode.AuthenticationRequired => StatusCodes.Status401Unauthorized,
RendezvousErrorCode.Forbidden => StatusCodes.Status403Forbidden, RendezvousErrorCode.Forbidden => StatusCodes.Status403Forbidden,
RendezvousErrorCode.NotFound or RendezvousErrorCode.StaleHost => StatusCodes.Status404NotFound, RendezvousErrorCode.NotFound => StatusCodes.Status404NotFound,
RendezvousErrorCode.Conflict or RendezvousErrorCode.ReplayRejected => StatusCodes.Status409Conflict, RendezvousErrorCode.Conflict
RendezvousErrorCode.Expired => StatusCodes.Status410Gone, or RendezvousErrorCode.IncompatibleProtocol
or RendezvousErrorCode.ReplayRejected => StatusCodes.Status409Conflict,
RendezvousErrorCode.Expired or RendezvousErrorCode.StaleHost => StatusCodes.Status410Gone,
RendezvousErrorCode.RateLimited or RendezvousErrorCode.CapacityExceeded => RendezvousErrorCode.RateLimited or RendezvousErrorCode.CapacityExceeded =>
StatusCodes.Status429TooManyRequests, StatusCodes.Status429TooManyRequests,
RendezvousErrorCode.ServiceUnavailable => StatusCodes.Status503ServiceUnavailable, RendezvousErrorCode.ServiceUnavailable => StatusCodes.Status503ServiceUnavailable,
@@ -347,7 +641,9 @@ internal static class ContractEndpoints
RendezvousErrorCode.NotFound => "The session was not found or is not owned by this publisher.", RendezvousErrorCode.NotFound => "The session was not found or is not owned by this publisher.",
RendezvousErrorCode.Conflict => "The session changed concurrently; retry with current state.", RendezvousErrorCode.Conflict => "The session changed concurrently; retry with current state.",
RendezvousErrorCode.Expired => "The session lease has expired.", RendezvousErrorCode.Expired => "The session lease has expired.",
RendezvousErrorCode.StaleHost => "The session has no fresh host presence.",
RendezvousErrorCode.IncompatibleProtocol => "The gameplay protocol is not enabled for this game.", RendezvousErrorCode.IncompatibleProtocol => "The gameplay protocol is not enabled for this game.",
RendezvousErrorCode.RateLimited => "The request rate limit was exceeded.",
RendezvousErrorCode.CapacityExceeded => "The configured session capacity is currently exhausted.", RendezvousErrorCode.CapacityExceeded => "The configured session capacity is currently exhausted.",
RendezvousErrorCode.ServiceUnavailable => "Session state is temporarily unavailable.", RendezvousErrorCode.ServiceUnavailable => "Session state is temporarily unavailable.",
RendezvousErrorCode.UnsupportedContractVersion => "The requested contract version is not supported.", RendezvousErrorCode.UnsupportedContractVersion => "The requested contract version is not supported.",
@@ -4,7 +4,8 @@ using Microsoft.AspNetCore.Diagnostics;
namespace FinalFactory.Rendezvous.Server.Http; namespace FinalFactory.Rendezvous.Server.Http;
internal sealed class RendezvousExceptionHandler : IExceptionHandler internal sealed class RendezvousExceptionHandler(
ILogger<RendezvousExceptionHandler> logger) : IExceptionHandler
{ {
public async ValueTask<bool> TryHandleAsync( public async ValueTask<bool> TryHandleAsync(
HttpContext httpContext, HttpContext httpContext,
@@ -17,16 +18,31 @@ internal sealed class RendezvousExceptionHandler : IExceptionHandler
} }
bool invalidRequest = exception is BadHttpRequestException or JsonException; bool invalidRequest = exception is BadHttpRequestException or JsonException;
httpContext.Response.StatusCode = invalidRequest bool payloadTooLarge = exception is BadHttpRequestException
{
StatusCode: StatusCodes.Status413PayloadTooLarge,
};
httpContext.Response.StatusCode = payloadTooLarge
? StatusCodes.Status413PayloadTooLarge
: invalidRequest
? StatusCodes.Status400BadRequest ? StatusCodes.Status400BadRequest
: StatusCodes.Status500InternalServerError; : StatusCodes.Status500InternalServerError;
LogRequestFailure(
logger,
payloadTooLarge ? "payload-too-large" : invalidRequest ? "invalid-request" : "internal-error",
httpContext.Response.StatusCode,
httpContext.Response.Headers["X-Rendezvous-Correlation-ID"].ToString() is { Length: > 0 } value
? value
: "unavailable");
await httpContext.Response.WriteAsJsonAsync( await httpContext.Response.WriteAsJsonAsync(
new ApiError new ApiError
{ {
Code = invalidRequest Code = invalidRequest
? RendezvousErrorCode.InvalidRequest ? RendezvousErrorCode.InvalidRequest
: RendezvousErrorCode.InternalError, : RendezvousErrorCode.InternalError,
Message = invalidRequest Message = payloadTooLarge
? "The request body exceeds the supported size."
: invalidRequest
? "The request body, route, or query value is invalid." ? "The request body, route, or query value is invalid."
: "The service could not complete the request.", : "The service could not complete the request.",
}, },
@@ -34,4 +50,16 @@ internal sealed class RendezvousExceptionHandler : IExceptionHandler
cancellationToken).ConfigureAwait(false); cancellationToken).ConfigureAwait(false);
return true; return true;
} }
private static readonly Action<ILogger, string, int, string, Exception?> RequestFailure =
LoggerMessage.Define<string, int, string>(
LogLevel.Warning,
new EventId(200, nameof(LogRequestFailure)),
"Request failed with {FailureKind} and HTTP status {StatusCode}; correlation {CorrelationId}");
private static void LogRequestFailure(
ILogger logger,
string failureKind,
int statusCode,
string correlationId) => RequestFailure(logger, failureKind, statusCode, correlationId, null);
} }
@@ -0,0 +1,96 @@
using System.Security.Cryptography;
using System.Text.Json;
using System.Text.Json.Serialization;
using FinalFactory.Rendezvous.Contracts;
using FinalFactory.Rendezvous.Server.Browser;
namespace FinalFactory.Rendezvous.Server.JoinAttempts;
internal sealed class JoinAttemptCursorCodec : IDisposable
{
private const string Prefix = "rvj1";
private readonly EphemeralCursorProtector _protector = new();
public string Encode(
SessionListingId listingId,
JoinAttemptId after,
DateTimeOffset now)
{
JoinAttemptCursorPayload payload = new()
{
ListingId = listingId.ToString(),
AfterAttemptId = after.ToString(),
ExpiresAtUnixSeconds = now.AddMinutes(5).ToUnixTimeSeconds(),
};
byte[] encoded = JsonSerializer.SerializeToUtf8Bytes(payload, ContractJson.Options);
try
{
return _protector.Protect(Prefix, encoded);
}
finally
{
CryptographicOperations.ZeroMemory(encoded);
}
}
public bool TryDecode(
string? cursor,
SessionListingId listingId,
DateTimeOffset now,
out JoinAttemptId? after)
{
after = null;
if (cursor is null)
{
return true;
}
if (!_protector.TryUnprotect(Prefix, cursor, out byte[] encodedPayload))
{
return false;
}
JoinAttemptCursorPayload? payload;
try
{
payload = JsonSerializer.Deserialize<JoinAttemptCursorPayload>(
encodedPayload,
ContractJson.Options);
}
catch (JsonException)
{
payload = null;
}
finally
{
CryptographicOperations.ZeroMemory(encodedPayload);
}
if (payload is null
|| payload.ExpiresAtUnixSeconds <= now.ToUnixTimeSeconds()
|| !string.Equals(payload.ListingId, listingId.ToString(), StringComparison.Ordinal)
|| !JoinAttemptId.TryParse(payload.AfterAttemptId, out JoinAttemptId attemptId))
{
return false;
}
after = attemptId;
return true;
}
public void Dispose() => _protector.Dispose();
public override string ToString() => "[JoinAttemptCursorCodec: key and cursors redacted]";
}
internal sealed class JoinAttemptCursorPayload
{
[JsonRequired]
public string ListingId { get; set; } = string.Empty;
[JsonRequired]
public string AfterAttemptId { get; set; } = string.Empty;
[JsonRequired]
public long ExpiresAtUnixSeconds { get; set; }
}
@@ -0,0 +1,343 @@
using System.Net;
using System.Security.Cryptography;
using System.Text.Json;
using FinalFactory.Rendezvous.Contracts;
using FinalFactory.Rendezvous.Server.Provisioning;
using FinalFactory.Rendezvous.Server.Sessions;
using FinalFactory.Rendezvous.Server.State;
namespace FinalFactory.Rendezvous.Server.JoinAttempts;
internal sealed record JoinAttemptServiceResult<T>(RendezvousErrorCode Error, T? Value = default)
{
public bool Succeeded => Error == RendezvousErrorCode.None;
}
internal sealed record ConnectionTicketGrant(string Ticket, DateTimeOffset ExpiresAt)
{
public override string ToString() => "[ConnectionTicketGrant: ticket redacted]";
}
internal sealed class JoinAttemptService(
GamePolicyRegistry policies,
IEphemeralRendezvousStore store,
ISessionCapabilityService capabilities,
JoinAttemptCursorCodec cursors,
IWallClock clock)
{
public string CreateAnonymousClientSubject(IPAddress remoteAddress)
{
ArgumentNullException.ThrowIfNull(remoteAddress);
IPAddress normalized = remoteAddress.IsIPv4MappedToIPv6
? remoteAddress.MapToIPv4()
: remoteAddress;
return capabilities.DeriveOpaqueIdentifier("join-http-client", normalized.ToString());
}
public JoinAttemptServiceResult<CreateJoinAttemptResponse> Create(
string clientSubject,
CreateJoinAttemptRequest request,
CancellationToken cancellationToken = default)
{
ArgumentNullException.ThrowIfNull(request);
if (string.IsNullOrWhiteSpace(clientSubject))
{
throw new ArgumentException("A bounded client subject is required.", nameof(clientSubject));
}
RendezvousErrorCode validation = ValidateCreate(request);
if (validation != RendezvousErrorCode.None)
{
return new(validation);
}
if (!policies.TryGet(request.GameId, request.EnvironmentId, out GamePolicy? policy)
|| policy is null)
{
return new(RendezvousErrorCode.NotFound);
}
if (!policy.AllowsProtocol(request.ProtocolVersion))
{
return new(RendezvousErrorCode.IncompatibleProtocol);
}
string requestFingerprint = ComputeRequestFingerprint(request);
string derivationSalt = capabilities.CreateDerivationSalt();
string hostCapability = Derive("join-host-punch", clientSubject, request, requestFingerprint, derivationSalt);
string clientCapability = Derive("join-client-punch", clientSubject, request, requestFingerprint, derivationSalt);
JoinAttemptId attemptId = new(capabilities.DeriveGuid(
"join-attempt-id",
clientSubject,
request.IdempotencyKey,
requestFingerprint,
derivationSalt));
string connectionTicket = NatIntroductionTokenCodec.Encode(
attemptId,
Derive("connection-ticket", clientSubject, request, requestFingerprint, derivationSalt));
if (!CredentialLengthsAreValid(hostCapability, clientCapability, connectionTicket)
|| !capabilities.TryFingerprint(hostCapability, out SecretFingerprint hostFingerprint)
|| !capabilities.TryFingerprint(clientCapability, out SecretFingerprint clientFingerprint)
|| !capabilities.TryFingerprint(connectionTicket, out SecretFingerprint ticketFingerprint))
{
throw new InvalidOperationException("Derived join credentials violated their contract invariants.");
}
MediationHandle mediationHandle = new(capabilities.DeriveGuid(
"join-mediation-handle",
clientSubject,
request.IdempotencyKey,
requestFingerprint,
derivationSalt));
StoreResult<StoredJoinAttempt> created = store.CreateJoinAttempt(new()
{
IdempotencyOwner = clientSubject,
IdempotencyKey = request.IdempotencyKey,
RequestFingerprint = requestFingerprint,
ClientSubject = clientSubject,
AttemptId = attemptId,
MediationHandle = mediationHandle,
Scope = new(request.GameId, request.EnvironmentId),
ListingId = request.ListingId,
ProtocolVersion = request.ProtocolVersion,
HostCapabilityFingerprint = hostFingerprint,
ClientCapabilityFingerprint = clientFingerprint,
ConnectionTicketFingerprint = ticketFingerprint,
CapabilityDerivationSalt = derivationSalt,
ScopeAttemptLimit = policy.MaxActiveJoinAttempts,
}, cancellationToken);
if (!created.Succeeded || created.Value is null)
{
return new(created.Code.ToContractError());
}
StoredJoinAttempt persisted = created.Value;
clientCapability = Derive(
"join-client-punch",
persisted.ClientSubject,
persisted.IdempotencyKey,
persisted.RequestFingerprint,
persisted.CapabilityDerivationSalt);
if (!capabilities.TryFingerprint(clientCapability, out SecretFingerprint persistedFingerprint)
|| persistedFingerprint != persisted.ClientCapabilityFingerprint)
{
throw new InvalidOperationException("Stored join state could not reproduce its client capability.");
}
return new(RendezvousErrorCode.None, new CreateJoinAttemptResponse
{
AttemptId = persisted.AttemptId,
MediationHandle = persisted.MediationHandle,
ClientPunchCapability = clientCapability,
ConnectionTicketDigest = NatIntroductionTokenCodec.ComputeDigest(
CreateConnectionTicket(persisted)),
ExpiresAt = persisted.ExpiresAt,
DedicatedFallback = StoredListing.CopyEndpoint(persisted.DedicatedFallback),
});
}
public JoinAttemptServiceResult<BrowseHostJoinAttemptsResponse> BrowseForHost(
SessionListingId listingId,
int contractVersion,
string? leaseToken,
int pageSize,
string? cursor,
CancellationToken cancellationToken = default)
{
RendezvousErrorCode version = ContractValidation.ValidateContractVersion(contractVersion);
if (version != RendezvousErrorCode.None)
{
return new(version);
}
if (!ContractValidation.IsOpaqueHttpCredentialValid(leaseToken)
|| !ContractValidation.IsPageSizeValid(pageSize)
|| !ContractValidation.IsCursorValid(cursor)
|| !capabilities.TryFingerprint(leaseToken, out SecretFingerprint leaseFingerprint))
{
return new(RendezvousErrorCode.InvalidRequest);
}
if (!cursors.TryDecode(cursor, listingId, clock.UtcNow, out JoinAttemptId? after))
{
return new(RendezvousErrorCode.InvalidRequest);
}
StoreResult<IReadOnlyList<StoredJoinAttempt>> found = store.BrowseHostJoinAttempts(new(
listingId,
leaseFingerprint,
pageSize + 1,
after), cancellationToken);
if (!found.Succeeded || found.Value is null)
{
return new(found.Code.ToContractError());
}
bool hasMore = found.Value.Count > pageSize;
StoredJoinAttempt[] page = found.Value.Take(pageSize).ToArray();
BrowseHostJoinAttemptsResponse response = new()
{
Items = page.Select(CreateHostAttempt).ToList(),
NextCursor = hasMore && page.Length > 0
? cursors.Encode(listingId, page[^1].AttemptId, clock.UtcNow)
: null,
};
int encodedBytes = JsonSerializer.SerializeToUtf8Bytes(response, ContractJson.Options).Length;
return ContractValidation.IsBrowserResponseSizeValid(encodedBytes)
? new(RendezvousErrorCode.None, response)
: new(RendezvousErrorCode.CapacityExceeded);
}
public JoinAttemptServiceResult<bool> Cancel(
JoinAttemptId attemptId,
string? clientPunchCapability,
CancellationToken cancellationToken = default)
{
if (!ContractValidation.IsCapabilityValid(clientPunchCapability)
|| !capabilities.TryFingerprint(clientPunchCapability, out SecretFingerprint fingerprint))
{
return new(RendezvousErrorCode.InvalidRequest);
}
StoreResult<bool> cancelled = store.CancelJoinAttempt(new(attemptId, fingerprint), cancellationToken);
return cancelled.Succeeded
? new(RendezvousErrorCode.None, true)
: new(cancelled.Code.ToContractError());
}
public JoinAttemptServiceResult<ConnectionTicketGrant> IssueConnectionTicket(
StoredJoinAttempt attempt)
{
ArgumentNullException.ThrowIfNull(attempt);
if (!attempt.IntroductionConsumed || attempt.IsCancelled)
{
return new(RendezvousErrorCode.Conflict);
}
if (attempt.ConnectionTicketExpiresAt <= clock.UtcNow)
{
return new(RendezvousErrorCode.Expired);
}
string ticket = CreateConnectionTicket(attempt);
if (!ContractValidation.IsConnectionTicketValid(ticket)
|| !capabilities.TryFingerprint(ticket, out SecretFingerprint fingerprint)
|| fingerprint != attempt.ConnectionTicketFingerprint)
{
throw new InvalidOperationException("Stored join state could not reproduce its connection ticket.");
}
return new(RendezvousErrorCode.None, new(ticket, attempt.ConnectionTicketExpiresAt));
}
private HostJoinAttempt CreateHostAttempt(StoredJoinAttempt attempt)
{
string capability = Derive(
"join-host-punch",
attempt.ClientSubject,
attempt.IdempotencyKey,
attempt.RequestFingerprint,
attempt.CapabilityDerivationSalt);
if (!ContractValidation.IsCapabilityValid(capability)
|| !capabilities.TryFingerprint(capability, out SecretFingerprint fingerprint)
|| fingerprint != attempt.HostCapabilityFingerprint)
{
throw new InvalidOperationException("Stored join state could not reproduce its host capability.");
}
return new()
{
AttemptId = attempt.AttemptId,
MediationHandle = attempt.MediationHandle,
HostPunchCapability = capability,
ConnectionTicketDigest = NatIntroductionTokenCodec.ComputeDigest(
CreateConnectionTicket(attempt)),
IsCancelled = attempt.IsCancelled,
ExpiresAt = attempt.ExpiresAt,
};
}
private string CreateConnectionTicket(StoredJoinAttempt attempt) =>
NatIntroductionTokenCodec.Encode(
attempt.AttemptId,
Derive(
"connection-ticket",
attempt.ClientSubject,
attempt.IdempotencyKey,
attempt.RequestFingerprint,
attempt.CapabilityDerivationSalt));
private static RendezvousErrorCode ValidateCreate(CreateJoinAttemptRequest request)
{
RendezvousErrorCode version = ContractValidation.ValidateContractVersion(request.ContractVersion);
if (version != RendezvousErrorCode.None)
{
return version;
}
return !ContractValidation.IsIdempotencyKeyValid(request.IdempotencyKey)
|| string.IsNullOrEmpty(request.GameId.Value)
|| string.IsNullOrEmpty(request.EnvironmentId.Value)
|| request.ListingId.Value == Guid.Empty
|| request.ProtocolVersion == 0
? RendezvousErrorCode.InvalidRequest
: RendezvousErrorCode.None;
}
private static string ComputeRequestFingerprint(CreateJoinAttemptRequest request)
{
byte[] encoded = JsonSerializer.SerializeToUtf8Bytes(request, ContractJson.Options);
byte[] digest = SHA256.HashData(encoded);
CryptographicOperations.ZeroMemory(encoded);
try
{
return Encode(digest);
}
finally
{
CryptographicOperations.ZeroMemory(digest);
}
}
private string Derive(
string purpose,
string clientSubject,
CreateJoinAttemptRequest request,
string requestFingerprint,
string derivationSalt) => Derive(
purpose,
clientSubject,
request.IdempotencyKey,
requestFingerprint,
derivationSalt);
private string Derive(
string purpose,
string clientSubject,
string idempotencyKey,
string requestFingerprint,
string derivationSalt) => capabilities.DeriveCapability(
purpose,
clientSubject,
idempotencyKey,
requestFingerprint,
derivationSalt);
private static bool CredentialLengthsAreValid(
string hostCapability,
string clientCapability,
string ticket) =>
ContractValidation.IsCapabilityValid(hostCapability)
&& ContractValidation.IsCapabilityValid(clientCapability)
&& ContractValidation.IsConnectionTicketValid(ticket)
&& hostCapability.Length == ContractLimits.DerivedCredentialCharacters
&& clientCapability.Length == ContractLimits.DerivedCredentialCharacters
&& ticket.Length == ContractLimits.DerivedCredentialCharacters
&& hostCapability.Length <= ContractLimits.LiteNetLibNatTokenMaxCharacters
&& clientCapability.Length <= ContractLimits.LiteNetLibNatTokenMaxCharacters;
private static string Encode(ReadOnlySpan<byte> bytes) => Convert
.ToBase64String(bytes)
.TrimEnd('=')
.Replace('+', '-')
.Replace('/', '_');
}
@@ -0,0 +1,14 @@
using System.ComponentModel.DataAnnotations;
namespace FinalFactory.Rendezvous.Server.Observability;
internal sealed class AuditOptions
{
public const string SectionName = "Rendezvous:Audit";
[Range(100, 100_000)]
public int MaxEntries { get; set; } = 10_000;
[Range(1, 30)]
public int RetentionDays { get; set; } = 30;
}
@@ -0,0 +1,142 @@
using System.Security.Cryptography;
using System.Text;
using Microsoft.Extensions.Options;
namespace FinalFactory.Rendezvous.Server.Observability;
internal sealed class AuditTrail
{
private readonly object _gate = new();
private readonly LinkedList<AuditEntry> _entries = [];
private readonly AuditOptions _options;
private readonly TimeProvider _timeProvider;
private readonly ILogger<AuditTrail> _logger;
private readonly RendezvousTelemetry _telemetry;
public AuditTrail(
IOptions<AuditOptions> options,
ILogger<AuditTrail> logger,
RendezvousTelemetry telemetry,
TimeProvider? timeProvider = null)
{
_options = options.Value;
_logger = logger;
_telemetry = telemetry;
_timeProvider = timeProvider ?? TimeProvider.System;
}
public void Record(
string actorSubject,
string action,
string result,
string targetKind,
string targetIdentifier,
string correlationId)
{
DateTimeOffset now = _timeProvider.GetUtcNow();
AuditEntry entry = new(
now,
Fingerprint(actorSubject),
action,
result,
targetKind,
Fingerprint(targetIdentifier),
correlationId);
lock (_gate)
{
PurgeExpired(now);
while (_entries.Count >= _options.MaxEntries)
{
_entries.RemoveFirst();
}
_entries.AddLast(entry);
}
_telemetry.RecordAudit(action, result);
LogOperatorAction(
_logger,
entry.ActorFingerprint,
action,
result,
targetKind,
entry.TargetFingerprint,
correlationId);
}
public IReadOnlyDictionary<string, long> GetAggregateCounts()
{
lock (_gate)
{
PurgeExpired(_timeProvider.GetUtcNow());
return _entries
.GroupBy(static entry => $"{entry.Action}:{entry.Result}", StringComparer.Ordinal)
.ToDictionary(
static group => group.Key,
static group => (long)group.Count(),
StringComparer.Ordinal);
}
}
internal IReadOnlyList<AuditEntry> GetEntriesForTests()
{
lock (_gate)
{
PurgeExpired(_timeProvider.GetUtcNow());
return _entries.ToArray();
}
}
private void PurgeExpired(DateTimeOffset now)
{
DateTimeOffset oldest = now.AddDays(-_options.RetentionDays);
while (_entries.First is { Value.Timestamp: var timestamp }
&& timestamp < oldest)
{
_entries.RemoveFirst();
}
}
private static string Fingerprint(string value)
{
byte[] digest = SHA256.HashData(Encoding.UTF8.GetBytes(value));
return Convert.ToHexString(digest.AsSpan(0, 12));
}
private static readonly Action<ILogger, string, string, string, string, string, string, Exception?>
OperatorAction = LoggerMessage.Define<string, string, string, string, string, string>(
LogLevel.Information,
new EventId(100, nameof(LogOperatorAction)),
"Operator audit: actor {ActorFingerprint} action {Action} completed with {Result} for {TargetKind} target {TargetFingerprint}; correlation {CorrelationId}");
private static void LogOperatorAction(
ILogger logger,
string actorFingerprint,
string action,
string result,
string targetKind,
string targetFingerprint,
string correlationId) => OperatorAction(
logger,
actorFingerprint,
action,
result,
targetKind,
targetFingerprint,
correlationId,
null);
}
internal sealed record AuditEntry(
DateTimeOffset Timestamp,
string ActorFingerprint,
string Action,
string Result,
string TargetKind,
string TargetFingerprint,
string CorrelationId)
{
public override string ToString() =>
$"[AuditEntry {Action}/{Result}; actor and target fingerprinted]";
}
@@ -0,0 +1,30 @@
using FinalFactory.Rendezvous.Contracts;
namespace FinalFactory.Rendezvous.Server.Observability;
internal static class HealthEndpoints
{
public static IEndpointRouteBuilder MapRendezvousHealthEndpoints(
this IEndpointRouteBuilder endpoints)
{
endpoints.MapGet(
"/health/live",
static () => Results.Ok(new HealthResponse { Status = "live" }))
.Produces<HealthResponse>()
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.WithName("GetLiveness")
.WithTags("Health");
endpoints.MapGet(
"/health/ready",
static (RendezvousReadiness readiness) =>
!readiness.GetSnapshot().IsReady
? Results.StatusCode(StatusCodes.Status503ServiceUnavailable)
: Results.Ok(new HealthResponse { Status = "ready" }))
.Produces<HealthResponse>()
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces(StatusCodes.Status503ServiceUnavailable)
.WithName("GetReadiness")
.WithTags("Health");
return endpoints;
}
}
@@ -0,0 +1,41 @@
using FinalFactory.Rendezvous.Server.Provisioning;
using FinalFactory.Rendezvous.Server.State;
using FinalFactory.Rendezvous.Server.Transport;
using Microsoft.Extensions.Options;
namespace FinalFactory.Rendezvous.Server.Observability;
internal sealed class RendezvousReadiness(
UdpMediatorService mediator,
ProvisioningReadiness provisioning,
IEphemeralRendezvousStore state,
IOptions<UdpMediatorOptions> udpOptions)
{
public ReadinessSnapshot GetSnapshot()
{
bool ipv6Required = !string.IsNullOrWhiteSpace(udpOptions.Value.Ipv6ListenAddress);
return new ReadinessSnapshot(
HttpListenerReady: true,
UdpIpv4ListenerReady: mediator.LocalEndpoint is not null,
UdpIpv6ListenerReady: !ipv6Required || mediator.LocalIpv6Endpoint is not null,
ProvisioningReady: provisioning.IsReady,
StoreAvailable: state.IsAvailable,
Draining: state.IsDraining);
}
}
internal sealed record ReadinessSnapshot(
bool HttpListenerReady,
bool UdpIpv4ListenerReady,
bool UdpIpv6ListenerReady,
bool ProvisioningReady,
bool StoreAvailable,
bool Draining)
{
public bool IsReady => HttpListenerReady
&& UdpIpv4ListenerReady
&& UdpIpv6ListenerReady
&& ProvisioningReady
&& StoreAvailable
&& !Draining;
}
@@ -0,0 +1,126 @@
using System.Diagnostics;
using System.Diagnostics.Metrics;
using FinalFactory.Rendezvous.Server.State;
namespace FinalFactory.Rendezvous.Server.Observability;
internal sealed class RendezvousTelemetry : IDisposable
{
public const string MeterName = "FinalFactory.Rendezvous";
public const string ActivitySourceName = "FinalFactory.Rendezvous.Server";
private readonly InMemoryEphemeralRendezvousStore _store;
private readonly Meter _meter = new(MeterName, "1.0.0");
private readonly ActivitySource _activities = new(ActivitySourceName, "1.0.0");
private readonly Counter<long> _httpRequests;
private readonly Histogram<double> _httpDuration;
private readonly Counter<long> _udpResults;
private readonly Histogram<double> _udpDuration;
private readonly Counter<long> _limiterDrops;
private readonly Counter<long> _auditEvents;
private readonly Counter<long> _connectionOutcomes;
private readonly Counter<long> _operatorAuthentication;
private readonly Histogram<double> _pairingLatency;
public RendezvousTelemetry(InMemoryEphemeralRendezvousStore store)
{
_store = store;
_httpRequests = _meter.CreateCounter<long>("rendezvous.http.requests");
_httpDuration = _meter.CreateHistogram<double>(
"rendezvous.http.duration",
"ms");
_udpResults = _meter.CreateCounter<long>("rendezvous.udp.results");
_udpDuration = _meter.CreateHistogram<double>(
"rendezvous.udp.duration",
"ms");
_limiterDrops = _meter.CreateCounter<long>("rendezvous.limiter.drops");
_auditEvents = _meter.CreateCounter<long>("rendezvous.audit.events");
_connectionOutcomes = _meter.CreateCounter<long>("rendezvous.connection.outcomes");
_operatorAuthentication = _meter.CreateCounter<long>("rendezvous.operator.authentication");
_pairingLatency = _meter.CreateHistogram<double>(
"rendezvous.pairing.latency",
"ms");
_meter.CreateObservableGauge(
"rendezvous.store.active_listings",
() => _store.GetMetricsSnapshot().ActiveListings);
_meter.CreateObservableGauge(
"rendezvous.store.active_leases",
() => _store.GetMetricsSnapshot().ActiveListings);
_meter.CreateObservableGauge(
"rendezvous.store.active_attempts",
() => _store.GetMetricsSnapshot().ActiveJoinAttempts);
_meter.CreateObservableGauge(
"rendezvous.queue.depth",
() => _store.GetMetricsSnapshot().ActiveJoinAttempts);
_meter.CreateObservableGauge(
"rendezvous.store.replay_markers",
() => _store.GetMetricsSnapshot().ReplayMarkers);
_meter.CreateObservableGauge(
"rendezvous.store.available",
() => _store.GetMetricsSnapshot().IsAvailable ? 1 : 0);
_meter.CreateObservableCounter(
"rendezvous.store.expiry_churn",
() => _store.GetMetricsSnapshot().ExpiryChurn);
}
public Activity? StartActivity(string name, ActivityKind kind = ActivityKind.Internal) =>
_activities.StartActivity(name, kind);
public void RecordHttp(string operation, int statusCode, double elapsedMilliseconds)
{
TagList tags = new()
{
{ "operation", operation },
{ "status_code", statusCode },
};
_httpRequests.Add(1, tags);
_httpDuration.Record(elapsedMilliseconds, tags);
}
public void RecordUdp(string operation, string result, double elapsedMilliseconds)
{
TagList tags = new()
{
{ "operation", operation },
{ "result", result },
};
_udpResults.Add(1, tags);
_udpDuration.Record(elapsedMilliseconds, tags);
}
public void RecordLimiterDrop(string transport, string partition) =>
_limiterDrops.Add(1, new TagList
{
{ "transport", transport },
{ "partition", partition },
});
public void RecordAudit(string action, string result) =>
_auditEvents.Add(1, new TagList
{
{ "action", action },
{ "result", result },
});
public void RecordConnectionOutcome(string outcome, string elapsedBucket) =>
_connectionOutcomes.Add(1, new TagList
{
{ "outcome", outcome },
{ "elapsed_bucket", elapsedBucket },
});
public void RecordOperatorAuthentication(string result) =>
_operatorAuthentication.Add(1, new TagList
{
{ "result", result },
});
public void RecordPairingLatency(double elapsedMilliseconds) =>
_pairingLatency.Record(elapsedMilliseconds);
public void Dispose()
{
_activities.Dispose();
_meter.Dispose();
}
}
@@ -0,0 +1,32 @@
using System.Diagnostics;
namespace FinalFactory.Rendezvous.Server.Observability;
internal sealed class TelemetryMiddleware(
RequestDelegate next,
RendezvousTelemetry telemetry)
{
public async Task InvokeAsync(HttpContext context)
{
string operation = context.GetEndpoint()?.Metadata.GetMetadata<IEndpointNameMetadata>()
?.EndpointName ?? "Unmatched";
long started = Stopwatch.GetTimestamp();
using Activity? activity = telemetry.StartActivity(
$"HTTP {operation}",
ActivityKind.Server);
string correlationId = activity?.TraceId.ToString() ?? Guid.NewGuid().ToString("N");
context.Response.Headers["X-Rendezvous-Correlation-ID"] = correlationId;
activity?.SetTag("rendezvous.operation", operation);
try
{
await next(context).ConfigureAwait(false);
}
finally
{
telemetry.RecordHttp(
operation,
context.Response.StatusCode,
Stopwatch.GetElapsedTime(started).TotalMilliseconds);
}
}
}
@@ -0,0 +1,428 @@
using FinalFactory.Rendezvous.Contracts;
using FinalFactory.Rendezvous.Server.Observability;
using FinalFactory.Rendezvous.Server.Provisioning;
using FinalFactory.Rendezvous.Server.State;
using Microsoft.AspNetCore.Mvc;
namespace FinalFactory.Rendezvous.Server.Operations;
internal static class OperatorEndpoints
{
private const string CorrelationHeader = "X-Rendezvous-Correlation-ID";
public static IEndpointRouteBuilder MapOperatorEndpoints(this IEndpointRouteBuilder endpoints)
{
RouteGroupBuilder group = endpoints.MapGroup("/v1/operator").WithTags("Operator");
group.MapGet("/status", GetStatus)
.Produces<OperatorStatusResponse>()
.Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.WithName("GetOperatorStatus");
group.MapPost("/listings/revoke", RevokeListing)
.Accepts<RevokeListingRequest>("application/json")
.Produces<OperatorActionResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("RevokeOperatorListing");
group.MapPost("/principals/revoke", RevokePrincipal)
.Accepts<RevokePrincipalRequest>("application/json")
.Produces<OperatorActionResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.Produces<ApiError>(StatusCodes.Status503ServiceUnavailable)
.WithName("RevokeOperatorPrincipal");
group.MapPost("/keys/revoke", RevokeSigningKey)
.Accepts<RevokeSigningKeyRequest>("application/json")
.Produces<OperatorActionResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.WithName("RevokeOperatorSigningKey");
group.MapPost("/drain", BeginDrain)
.Accepts<BeginDrainRequest>("application/json")
.Produces<OperatorActionResponse>()
.Produces<ApiError>(StatusCodes.Status400BadRequest)
.Produces<ApiError>(StatusCodes.Status413PayloadTooLarge)
.Produces<ApiError>(StatusCodes.Status401Unauthorized)
.Produces<ApiError>(StatusCodes.Status403Forbidden)
.Produces<ApiError>(StatusCodes.Status404NotFound)
.Produces<ApiError>(StatusCodes.Status429TooManyRequests)
.WithName("BeginOperatorDrain");
return endpoints;
}
private static IResult GetStatus(
[FromHeader(Name = "Authorization")] string? authorization,
[FromServices] PrincipalCredentialService credentials,
[FromServices] IWallClock clock,
[FromServices] OperatorService service,
[FromServices] AuditTrail audit,
[FromServices] RendezvousTelemetry telemetry,
HttpContext context)
{
if (!TryAuthorize(
authorization,
OperatorPermission.ReadPolicy,
"inspect-status",
credentials,
clock,
audit,
telemetry,
context,
out OperatorPrincipal? principal,
out IResult? failure))
{
return failure!;
}
OperatorStatusResponse response = service.GetStatus();
audit.Record(
principal!.Subject,
"inspect-status",
"succeeded",
"service",
"rendezvous",
Correlation(context));
return Results.Ok(response);
}
private static IResult RevokeListing(
[FromBody] RevokeListingRequest request,
[FromHeader(Name = "Authorization")] string? authorization,
[FromServices] PrincipalCredentialService credentials,
[FromServices] IWallClock clock,
[FromServices] OperatorService service,
[FromServices] AuditTrail audit,
[FromServices] RendezvousTelemetry telemetry,
HttpContext context,
CancellationToken cancellationToken)
{
if (!TryAuthorize(
authorization,
OperatorPermission.RevokePublisher,
"revoke-listing",
credentials,
clock,
audit,
telemetry,
context,
out OperatorPrincipal? principal,
out IResult? failure))
{
return failure!;
}
bool valid = SessionListingId.TryParse(request.ListingId, out SessionListingId listingId);
if (!valid || !string.Equals(request.ListingId, request.ConfirmListingId, StringComparison.Ordinal))
{
AuditRejected(audit, principal!, "revoke-listing", "listing", request.ListingId, context);
return BadRequest("A valid listing ID and an exact repeated confirmation are required.");
}
StoreResult<bool> result = service.RevokeListing(listingId, cancellationToken);
return StoreActionResult(
result.Code,
audit,
principal!,
"revoke-listing",
"listing",
request.ListingId,
context,
affectedResources: result.Succeeded ? 1 : null);
}
private static IResult RevokePrincipal(
[FromBody] RevokePrincipalRequest request,
[FromHeader(Name = "Authorization")] string? authorization,
[FromServices] PrincipalCredentialService credentials,
[FromServices] IWallClock clock,
[FromServices] OperatorService service,
[FromServices] AuditTrail audit,
[FromServices] RendezvousTelemetry telemetry,
HttpContext context,
CancellationToken cancellationToken)
{
if (!TryAuthorize(
authorization,
OperatorPermission.RevokePublisher,
"revoke-principal",
credentials,
clock,
audit,
telemetry,
context,
out OperatorPrincipal? principal,
out IResult? failure))
{
return failure!;
}
bool safeSubject = request.Subject is { Length: > 0 and <= 128 }
&& request.Subject.All(static character => character is >= '!' and <= '~');
if (!safeSubject
|| !string.Equals(request.Subject, request.ConfirmSubject, StringComparison.Ordinal)
|| request.LifetimeSeconds is < 1 or > 600)
{
AuditRejected(audit, principal!, "revoke-principal", "principal", request.Subject, context);
return BadRequest("A valid subject, exact repeated confirmation, and 1-600 second lifetime are required.");
}
StoreResult<int> result = service.RevokePrincipal(
request.Subject,
TimeSpan.FromSeconds(request.LifetimeSeconds),
cancellationToken);
return StoreActionResult(
result.Code,
audit,
principal!,
"revoke-principal",
"principal",
request.Subject,
context,
result.Value);
}
private static IResult RevokeSigningKey(
[FromBody] RevokeSigningKeyRequest request,
[FromHeader(Name = "Authorization")] string? authorization,
[FromServices] PrincipalCredentialService credentials,
[FromServices] IWallClock clock,
[FromServices] OperatorService service,
[FromServices] AuditTrail audit,
[FromServices] RendezvousTelemetry telemetry,
HttpContext context)
{
if (!TryAuthorize(
authorization,
OperatorPermission.RotateKeys,
"revoke-signing-key",
credentials,
clock,
audit,
telemetry,
context,
out OperatorPrincipal? principal,
out IResult? failure))
{
return failure!;
}
bool safeKeyId = request.KeyId is { Length: > 0 and <= 64 }
&& request.KeyId.All(static character => character is
>= 'A' and <= 'Z'
or >= 'a' and <= 'z'
or >= '0' and <= '9'
or '-'
or '_');
if (!safeKeyId || !string.Equals(request.KeyId, request.ConfirmKeyId, StringComparison.Ordinal))
{
AuditRejected(audit, principal!, "revoke-signing-key", "signing-key", request.KeyId, context);
return BadRequest("A valid key ID and an exact repeated confirmation are required.");
}
bool revoked = service.RevokeSigningKey(request.KeyId);
string result = revoked ? "succeeded" : "not-found";
audit.Record(
principal!.Subject,
"revoke-signing-key",
result,
"signing-key",
request.KeyId,
Correlation(context));
return revoked
? Results.Ok(new OperatorActionResponse { Status = "completed" })
: Error(RendezvousErrorCode.NotFound, "The requested resource was not found.");
}
private static IResult BeginDrain(
[FromBody] BeginDrainRequest request,
[FromHeader(Name = "Authorization")] string? authorization,
[FromServices] PrincipalCredentialService credentials,
[FromServices] IWallClock clock,
[FromServices] OperatorService service,
[FromServices] AuditTrail audit,
[FromServices] RendezvousTelemetry telemetry,
HttpContext context,
CancellationToken cancellationToken)
{
if (!TryAuthorize(
authorization,
OperatorPermission.ManagePolicy,
"begin-drain",
credentials,
clock,
audit,
telemetry,
context,
out OperatorPrincipal? principal,
out IResult? failure))
{
return failure!;
}
if (!string.Equals(request.Confirmation, "DRAIN", StringComparison.Ordinal))
{
AuditRejected(audit, principal!, "begin-drain", "service", "rendezvous", context);
return BadRequest("The confirmation value must be exactly 'DRAIN'.");
}
service.BeginDrain(cancellationToken);
audit.Record(
principal!.Subject,
"begin-drain",
"succeeded",
"service",
"rendezvous",
Correlation(context));
return Results.Ok(new OperatorActionResponse { Status = "draining" });
}
private static bool TryAuthorize(
string? authorization,
OperatorPermission requiredPermission,
string operation,
PrincipalCredentialService credentials,
IWallClock clock,
AuditTrail audit,
RendezvousTelemetry telemetry,
HttpContext context,
out OperatorPrincipal? principal,
out IResult? failure)
{
principal = null;
failure = null;
const string prefix = "Bearer ";
if (authorization is null
|| !authorization.StartsWith(prefix, StringComparison.OrdinalIgnoreCase))
{
telemetry.RecordOperatorAuthentication("rejected");
failure = AuthenticationRequired(context);
return false;
}
CredentialValidationResult validation = credentials.Validate(
authorization[prefix.Length..],
clock.UtcNow);
if (!validation.IsValid || validation.Principal is not OperatorPrincipal candidate)
{
telemetry.RecordOperatorAuthentication("rejected");
failure = AuthenticationRequired(context);
return false;
}
if (!candidate.Permissions.Contains(requiredPermission))
{
telemetry.RecordOperatorAuthentication("forbidden");
audit.Record(
candidate.Subject,
operation,
"forbidden",
"operator-operation",
operation,
Correlation(context));
failure = Error(RendezvousErrorCode.Forbidden, "The operator is not authorized for this operation.");
return false;
}
telemetry.RecordOperatorAuthentication("accepted");
principal = candidate;
return true;
}
private static IResult StoreActionResult(
StoreResultCode code,
AuditTrail audit,
OperatorPrincipal principal,
string action,
string targetKind,
string targetIdentifier,
HttpContext context,
int? affectedResources)
{
string auditResult = code == StoreResultCode.Success
? "succeeded"
: code.ToString().ToLowerInvariant();
audit.Record(
principal.Subject,
action,
auditResult,
targetKind,
targetIdentifier,
Correlation(context));
return code switch
{
StoreResultCode.Success => Results.Ok(new OperatorActionResponse
{
Status = "completed",
AffectedResources = affectedResources,
}),
StoreResultCode.NotFound => Error(
RendezvousErrorCode.NotFound,
"The requested resource was not found."),
StoreResultCode.CapacityExceeded => Error(
RendezvousErrorCode.CapacityExceeded,
"The operation could not be retained within the configured capacity."),
StoreResultCode.ServiceUnavailable or StoreResultCode.Draining => Error(
RendezvousErrorCode.ServiceUnavailable,
"The service is not available for this operation."),
_ => Error(RendezvousErrorCode.Conflict, "The operation could not be completed."),
};
}
private static void AuditRejected(
AuditTrail audit,
OperatorPrincipal principal,
string action,
string targetKind,
string? targetIdentifier,
HttpContext context) => audit.Record(
principal.Subject,
action,
"rejected",
targetKind,
targetIdentifier ?? string.Empty,
Correlation(context));
private static string Correlation(HttpContext context) =>
context.Response.Headers[CorrelationHeader].ToString() is { Length: > 0 } value
? value
: "unavailable";
private static IResult AuthenticationRequired(HttpContext context)
{
context.Response.Headers.WWWAuthenticate = "Bearer realm=\"operator\"";
return Error(
RendezvousErrorCode.AuthenticationRequired,
"A valid operator bearer credential is required.");
}
private static IResult BadRequest(string message) => Error(RendezvousErrorCode.InvalidRequest, message);
private static IResult Error(RendezvousErrorCode code, string message) => Results.Json(
new ApiError { Code = code, Message = message },
ContractJson.Options,
statusCode: code switch
{
RendezvousErrorCode.AuthenticationRequired => StatusCodes.Status401Unauthorized,
RendezvousErrorCode.Forbidden => StatusCodes.Status403Forbidden,
RendezvousErrorCode.NotFound => StatusCodes.Status404NotFound,
RendezvousErrorCode.CapacityExceeded => StatusCodes.Status429TooManyRequests,
RendezvousErrorCode.ServiceUnavailable => StatusCodes.Status503ServiceUnavailable,
RendezvousErrorCode.Conflict => StatusCodes.Status409Conflict,
_ => StatusCodes.Status400BadRequest,
});
}
@@ -0,0 +1,95 @@
namespace FinalFactory.Rendezvous.Server.Operations;
internal sealed record OperatorStatusResponse
{
public required string Status { get; init; }
public required OperatorCompatibilityResponse Compatibility { get; init; }
public required OperatorReadinessResponse Readiness { get; init; }
public required OperatorStoreResponse Store { get; init; }
public required IReadOnlyList<OperatorTenantResponse> Tenants { get; init; }
public required IReadOnlyList<OperatorSigningKeyResponse> SigningKeys { get; init; }
public required IReadOnlyDictionary<string, long> AuditCounts { get; init; }
}
internal sealed record OperatorCompatibilityResponse
{
public required string ServerVersion { get; init; }
public required string MinimumClientVersion { get; init; }
public required int MaximumClientMajorVersion { get; init; }
public required IReadOnlyList<int> HttpContractVersions { get; init; }
public required IReadOnlyList<int> UdpContractVersions { get; init; }
public required IReadOnlyList<int> ConnectionTicketFormatVersions { get; init; }
public required int LiteNetLibMajorVersion { get; init; }
public required string GameplayProtocolCompatibility { get; init; }
}
internal sealed record OperatorReadinessResponse
{
public required bool HttpListener { get; init; }
public required bool UdpIpv4Listener { get; init; }
public required bool UdpIpv6Listener { get; init; }
public required bool Provisioning { get; init; }
public required bool Store { get; init; }
public required bool Draining { get; init; }
}
internal sealed record OperatorStoreResponse
{
public required int ActiveListings { get; init; }
public required int FreshPresenceBindings { get; init; }
public required int ActiveJoinAttempts { get; init; }
public required int RetainedOutcomeReports { get; init; }
public required int ReplayMarkers { get; init; }
public required int PrincipalRevocations { get; init; }
public required int IdempotencyEntries { get; init; }
public required long MaintenanceSweeps { get; init; }
public required long ExpiryChurn { get; init; }
}
internal sealed record OperatorTenantResponse
{
public required string GameId { get; init; }
public required string EnvironmentId { get; init; }
public required string Status { get; init; }
}
internal sealed record OperatorSigningKeyResponse
{
public required string KeyId { get; init; }
public required string Status { get; init; }
public required DateTimeOffset SignUntil { get; init; }
public required DateTimeOffset VerifyUntil { get; init; }
public string? GameId { get; init; }
public string? EnvironmentId { get; init; }
public required IReadOnlyList<string> CredentialKinds { get; init; }
}
internal sealed record OperatorActionResponse
{
public required string Status { get; init; }
public int? AffectedResources { get; init; }
}
internal sealed record RevokeListingRequest
{
public required string ListingId { get; init; }
public required string ConfirmListingId { get; init; }
}
internal sealed record RevokePrincipalRequest
{
public required string Subject { get; init; }
public required string ConfirmSubject { get; init; }
public required int LifetimeSeconds { get; init; }
}
internal sealed record RevokeSigningKeyRequest
{
public required string KeyId { get; init; }
public required string ConfirmKeyId { get; init; }
}
internal sealed record BeginDrainRequest
{
public required string Confirmation { get; init; }
}

Some files were not shown because too many files have changed in this diff Show More