Fix CI build/deploy/smoke pipeline for the self-hosted qa runner (#3)
All checks were successful
CI / build-and-push (push) Successful in 39s
CI / deploy-qa (push) Successful in 12s
CI / smoke-qa (push) Successful in 18s

## Summary
- Fix a chain of QA CI issues: Docker build/health-check flakiness, NuGet vuln pins, then adds the post-deploy integration + Playwright smoke suite and works through everything needed to make it actually run on the self-hosted `qa` runner (musl/Alpine job container, no node/dotnet/curl preinstalled, docker-outside-of-docker networking).
- Adds a fixed dev/QA seed admin user + fixed Development environment API key so integration tests and e2e specs have a stable target.
- Pins Aspire's `AppHost.cs` ports/credentials to match the docker-compose local dev defaults.
- Adds `tests/api/MicCheck.Api.Tests.Integration` coverage for disabled flags, environment-document bootstrap, identity override precedence, auth login, and unauthorized access; adds a Playwright e2e suite under `src/admin/e2e` (login, nav, context selection, features CRUD/toggle).
- Adds a `smoke-qa` CI job that runs both suites against the just-deployed QA stack, working around: no curl/node/dotnet on the bare runner, musl vs glibc (Playwright browsers run via the official `mcr.microsoft.com/playwright` image instead), and the runner's job-container network isolation (reach the QA stack via the docker bridge gateway IP; `docker cp` instead of a bind mount to get files into the playwright container, since paths don't cross the docker-outside-of-docker boundary).

## Test plan
- [x] Unit tests pass (dotnet test tests/api/MicCheck.Api.Tests.Unit, 243 passed)
- [x] API integration suite passes against the real QA stack in CI
- [x] Playwright e2e suite passes against the real QA stack in CI (verified 3/3 locally against a real dev API + vite server for the flakiest spec)
- [x] Full CI pipeline (build → deploy-qa → smoke-qa) green end to end on the qa runner

Reviewed-on: #3
This commit was merged in pull request #3.
This commit is contained in:
2026-07-04 18:53:05 -07:00
parent e10cba77ed
commit 1556b486d2
39 changed files with 1344 additions and 14 deletions

View File

@@ -37,6 +37,104 @@ image_names() {
ADMIN_IMAGE="$REGISTRY/$REGISTRY_OWNER/miccheck-admin"
}
# Installs .NET into $CI_ROOT/.dotnet via the vendored dotnet-install.sh if
# `dotnet` isn't already on PATH, then prepends it to PATH for this process.
# Keeps bare runners (no SDK preinstalled) working the same as a dev machine.
ensure_dotnet() {
if command -v dotnet > /dev/null 2>&1; then
return 0
fi
local install_dir="$CI_ROOT/.dotnet"
if [[ ! -x "$install_dir/dotnet" ]]; then
log "dotnet not found on PATH; installing .NET SDK via dotnet-install.sh"
bash "$CI_ROOT/dotnet-install.sh" --channel LTS --install-dir "$install_dir"
fi
export PATH="$install_dir:$PATH"
export DOTNET_ROOT="$install_dir"
ensure_dotnet_native_deps
}
# The .NET runtime is a native ELF binary that dynamically links libstdc++/libgcc.
# Bare/minimal images (e.g. the act hostexecutor container) may lack them entirely,
# which fails as an obscure symbol-relocation error rather than "command not found".
ensure_dotnet_native_deps() {
if ldconfig -p 2>/dev/null | grep -q 'libstdc++\.so\.6'; then
return 0
fi
log "libstdc++.so.6 missing; installing native runtime deps for dotnet"
local sudo_cmd=""
if [[ "$(id -u)" -ne 0 ]] && command -v sudo > /dev/null 2>&1; then
sudo_cmd="sudo"
fi
if command -v apt-get > /dev/null 2>&1; then
$sudo_cmd apt-get update -y
$sudo_cmd apt-get install -y --no-install-recommends libstdc++6 libgcc-s1 libicu-dev ca-certificates
elif command -v apk > /dev/null 2>&1; then
# Alpine/musl runner - dotnet-install.sh falls back to the linux-musl-x64 SDK
# here, which still wants a real libstdc++/libgcc (not just gcompat).
$sudo_cmd apk add --no-cache libstdc++ libgcc icu-libs ca-certificates
else
fail "libstdc++.so.6 missing and neither apt-get nor apk is available; install a C++ runtime manually on this runner"
fi
}
# Installs Node.js into $CI_ROOT/.node from the official prebuilt tarball if
# `npm` isn't already on PATH, then prepends it to PATH for this process.
# Mirrors ensure_dotnet() above - keeps bare runners (no Node preinstalled,
# e.g. the self-hosted qa runner) working the same as a dev machine.
ensure_node() {
if command -v npm > /dev/null 2>&1; then
return 0
fi
# nodejs.org only ships glibc binaries; on a musl/Alpine runner (same one
# dotnet-install.sh detects and picks the linux-musl-x64 SDK for) that
# tarball fails to exec at all ("env: can't execute 'node'"). Prefer the
# distro's own package on musl instead of a broken glibc download.
if [[ ! -x "$CI_ROOT/.node/bin/node" ]] && command -v apk > /dev/null 2>&1; then
log "node not found on PATH; installing via apk (musl runner)"
local sudo_cmd=""
if [[ "$(id -u)" -ne 0 ]] && command -v sudo > /dev/null 2>&1; then
sudo_cmd="sudo"
fi
$sudo_cmd apk add --no-cache nodejs npm
return 0
fi
local node_version="22.14.0"
local install_dir="$CI_ROOT/.node"
if [[ ! -x "$install_dir/bin/node" ]]; then
log "node not found on PATH; installing Node.js v$node_version"
local tarball="node-v${node_version}-linux-x64"
local url="https://nodejs.org/dist/v${node_version}/${tarball}.tar.xz"
if command -v curl > /dev/null 2>&1; then
curl -fsSL "$url" -o "/tmp/${tarball}.tar.xz"
elif command -v wget > /dev/null 2>&1; then
wget -q "$url" -O "/tmp/${tarball}.tar.xz"
else
fail "neither curl nor wget found on PATH; cannot download Node.js"
fi
mkdir -p "$install_dir"
tar -xJf "/tmp/${tarball}.tar.xz" -C "$install_dir" --strip-components=1
rm -f "/tmp/${tarball}.tar.xz"
fi
export PATH="$install_dir/bin:$PATH"
}
# The IP address on which a container published on 0.0.0.0/<gateway-ip> is
# reachable from a sibling container on docker's default bridge network (i.e.
# the docker host's bridge-side address, not its public interface). Used to
# bind QA's db port narrowly - reachable by the smoke-qa job container, not
# exposed off-box the way 0.0.0.0 would be.
docker_bridge_gateway() {
docker network inspect bridge -f '{{(index .IPAM.Config 0).Gateway}}' 2>/dev/null \
|| ip route show default 2>/dev/null | awk '/default/ {print $3; exit}'
}
registry_login() {
require_registry_vars
[[ -n "$REGISTRY_USER" ]] || fail "REGISTRY_USER env var is required to push images"