Files
certctl/.github/workflows/ci.yml
T
shankar0123 86fffa305a fix(deploy,helm,docs): published-image HEALTHCHECK speaks HTTPS + Helm /ready path + docs HTTPS sweep (U-2)
Pre-U-2 the published `ghcr.io/shankar0123/certctl-server` image
shipped with `HEALTHCHECK CMD curl -f http://localhost:8443/health`.
The server has been HTTPS-only since the v2.2 HTTPS-Everywhere milestone
(`cmd/server/main.go::ListenAndServeTLS`, no plaintext fallback, TLS
1.3 pinned), so the probe failed on every interval and Docker marked
the container `unhealthy` indefinitely. Operators inside docker-
compose / Helm / the example stacks were unaffected — compose overrides
the HEALTHCHECK with `--cacert + https://`, Helm uses explicit
`httpGet` probes that ignore Docker's HEALTHCHECK, and every example
compose file overrides with `curl -sfk https://localhost:8443/health`.
But anyone running bare `docker run` / Docker Swarm / Nomad / ECS —
exactly the "I just pulled the published image" path — saw permanent
`unhealthy` status and (depending on orchestrator policy) a restart-
loop. (Audit: cat-u-healthcheck_protocol_mismatch in
coverage-gap-audit-2026-04-24-v5/unified-audit.md.)

Recon for U-2 surfaced two adjacent bugs from the same v2.2 milestone
gap, both bundled into this commit because they share the same root
cause and the same operator surface:

  1. Helm chart `server.readinessProbe.httpGet.path` pointed at
     `/readyz`, the kube-flavored convention. The certctl server
     doesn't register `/readyz` (only `/health` and `/ready` are
     wired and bypass the auth middleware — see
     internal/api/router/router.go:81 and cmd/server/main.go:920).
     K8s readiness probes therefore got 401 (api-key auth rejection)
     or 404 (when auth was disabled), pods stayed `NotReady`
     indefinitely, and Helm rollouts stalled.

  2. The agent image (`Dockerfile.agent`) had no HEALTHCHECK at all,
     so bare-`docker run` agents got zero health signal. The
     compose override at `deploy/docker-compose.yml:173` called
     `pgrep -f certctl-agent` against the agent image, but the
     agent image didn't ship `procps` — pgrep was missing too. The
     compose probe was a latent always-fail.

We fixed all three with the audit-recommended shape (option (a) — `-k`)
plus three structural backstops:

Files changed:

Phase 1 — Dockerfile fix:
- Dockerfile: HEALTHCHECK switched from `curl -f http://localhost:8443/
  health` to `curl -fsk https://localhost:8443/health`. `-k`
  (insecure) is acceptable because the probe is localhost-to-localhost:
  the same process serving the cert is being probed, no network hop.
  Pinning `--cacert` is not viable for the published image because
  the bootstrap cert is per-deploy (generated into the `certs` named
  volume on first up; operator-supplied via Helm's `existingSecret`
  or cert-manager). Long-form docblock cross-references the audit
  closure, the compose vs Helm vs examples coverage matrix, and the
  CI guardrail.
- Dockerfile.agent: added HEALTHCHECK using `pgrep -f certctl-agent`
  matching the compose pattern. Added `procps` to the runtime apk
  install — fixes both the new image-level HEALTHCHECK AND the
  pre-existing compose probe that was silently failing.

Phase 2 — Helm readiness probe path:
- deploy/helm/certctl/values.yaml: server.readinessProbe.httpGet.path
  changed from `/readyz` to `/ready`. Liveness probe path
  (`/health`) was correct and is unchanged. Probes block now carries
  an explanatory comment naming the registered no-auth probe routes
  and the U-2 closure rationale.

Phase 3 — Image-level integration tests:
- deploy/test/healthcheck_test.go (new, //go:build integration):
  TestPublishedServerImage_HealthcheckSpecUsesHTTPS builds the server
  image, inspects `Config.Healthcheck.Test` via `docker inspect`,
  and asserts the array contains `https://localhost:8443/health` and
  `-k`, and does NOT contain `http://localhost:8443/health`
  (positive + negative regression contracts).
  TestPublishedAgentImage_HealthcheckSpecExists builds the agent image
  and asserts the HEALTHCHECK uses `pgrep` against `certctl-agent`.
  Both tests `t.Skip` cleanly when docker isn't available (sandbox /
  CI without docker-in-docker) — verified locally: tests skip with the
  diagnostic and the suite returns PASS.
  TestPublishedServerImage_HealthcheckTransitionsToHealthy is a
  documented `t.Skip` placeholder until the harness wires a sidecar
  postgres for image-level smoke; the spec-level tests above cover the
  audit-flagged regression.

Phase 4 — CI guardrail:
- .github/workflows/ci.yml: new "Forbidden plaintext HEALTHCHECK
  regression guard (U-2)" step. Scoped patterns catch
  `HEALTHCHECK.*http://` and `curl -f http://localhost:8443/health`
  in any `Dockerfile*`. Comment lines exempt; docs/upgrade-to-tls.md
  out of scope (the post-cutover invariant string at line 182 is
  intentionally a documented expected-failure assertion). Verified
  locally on the real tree (passes) and against synthetic regressions
  (each fires the guard).

Phase 5 — Docs sweep:
- docs/connectors.md: 15 stale curl examples updated from
  `http://localhost:8443/...` to `https://localhost:8443/...` with
  `--cacert "$CA"` injected on every site. Added a one-time
  introductory note documenting the `$CA` extraction with
  `docker compose ... exec ... cat /etc/certctl/tls/ca.crt`,
  matching the pattern in docs/quickstart.md. Pre-U-2 these examples
  silently failed against the HTTPS listener.

Phase 6 — Release surface:
- CHANGELOG.md: appended U-2 section to the existing [unreleased]
  block (immediately below the G-1 entry). Sections: explanatory
  blockquote covering all three bugs (primary + 2 adjacent), Fixed,
  Added, Changed.

Verification (all gates pass):
- go build ./... — clean
- go vet ./... — clean
- go vet -tags integration ./deploy/test/ — clean
- go test -short ./... — every package green
- go test -tags integration -v -run TestPublishedServerImage|TestPublishedAgentImage ./deploy/test/ —
  three tests SKIP cleanly with "docker not available" diagnostic
- helm lint deploy/helm/certctl/ — clean
- helm template smoke render — succeeds; rendered Deployment carries
  `path: /ready` and zero `/readyz` matches
- python3 yaml.safe_load on api/openapi.yaml — parses
- govulncheck ./... — no vulnerabilities in our code
- CI guardrail mirror: clean on real tree, fires on synthetic
  regression patterns

Out of scope (intentionally untouched):
- cmd/server/main.go::ListenAndServeTLS — HTTPS-only is correct,
  this finding does NOT propose adding back a plaintext listener.
- deploy/docker-compose.yml:126 HEALTHCHECK — already correct.
- deploy/docker-compose.test.yml HEALTHCHECK blocks — already correct.
- All 5 examples/*/docker-compose.yml HEALTHCHECK overrides — already
  correct (they ALSO use `-fsk https://localhost:8443/health`).
- Helm server.livenessProbe.httpGet — already uses `scheme: HTTPS` +
  `path: /health`, correct.
- docs/upgrade-to-tls.md:182 `curl ... http://localhost:8443/health`
  invariant line — that's the expected-failure assertion for the
  post-cutover state ("plaintext is gone, expect Connection refused");
  intentionally left intact.
- Go production code — this is purely a deploy-image / probe / docs /
  Helm-chart fix.

Refs: coverage-gap-audit-2026-04-24-v5/unified-audit.md
      §2 P1 cluster, cat-u-healthcheck_protocol_mismatch
      Audit recommendation followed verbatim: 'change Dockerfile:80
      to CMD curl -kf https://localhost:8443/health'.
2026-04-25 12:02:18 +00:00

351 lines
15 KiB
YAML

name: CI
on:
push:
branches:
- master
- v2-dev
pull_request:
branches:
- master
jobs:
go-build-and-test:
name: Go Build & Test
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Go
uses: actions/setup-go@v5
with:
go-version: '1.25.9'
- name: Go Build
run: |
go build ./cmd/server/...
go build ./cmd/agent/...
go build ./cmd/mcp-server/...
go build ./cmd/cli/...
- name: Go Vet
run: go vet ./...
- name: Install golangci-lint
run: |
curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(go env GOPATH)/bin v2.11.4
- name: Run golangci-lint
run: golangci-lint run ./... --timeout 5m
- name: Install govulncheck
run: go install golang.org/x/vuln/cmd/govulncheck@latest
- name: Run govulncheck
run: govulncheck ./...
- name: Forbidden auth-type literal regression guard (G-1)
# G-1 closed the JWT silent auth downgrade by removing "jwt" from the
# accepted CERTCTL_AUTH_TYPE values. This step grep-fails the build
# if "jwt" reappears in any of the *additive* auth-type surfaces:
# the validAuthTypes / ValidAuthTypes() set, the OpenAPI enum, the
# helm chart's allowed-types list, or the .env.example default.
# Comment lines and the dedicated rejection branch in config.go
# (`c.Auth.Type == "jwt"`) are intentionally exempt — those are the
# G-1 fix itself, not a regression.
#
# Connector packages (internal/connector/) are exempt because the
# Google OAuth2 service-account JWT and step-ca provisioner one-
# time-token JWT are external-protocol uses, unrelated to certctl's
# own auth shape. Test files (_test.go) are exempt so negative
# tests can pass the literal.
#
# See docs/upgrade-to-v2-jwt-removal.md for the closure rationale,
# or internal/config/config.go::ValidAuthTypes for the allowed set.
run: |
set -e
# Scoped patterns that indicate "jwt" being added back to an
# allowed-set surface. Each catches a regression shape we've
# actually seen in pre-G-1 code:
# - Go map/slice literal: "jwt": true or "jwt",
# - Go switch case: case "jwt"
# - YAML enum: enum: [..., jwt, ...] or - jwt
# - .env conditional: AUTH_TYPE.*"jwt"|=jwt$
BAD=$(grep -rnEH \
-e '"jwt"\s*:\s*true' \
-e '"jwt"\s*,' \
-e 'case\s+"jwt"' \
-e 'enum:.*\bjwt\b' \
-e '^\s*-\s*jwt\s*$' \
-e 'AUTH_TYPE\s*=\s*jwt\s*$' \
-e 'AUTH_TYPE\s*=\s*jwt\s*#' \
-e 'auth\.type\s*=\s*jwt\s*$' \
-e 'AuthType\("jwt"\)' \
internal/config/ \
internal/api/ \
cmd/ \
api/openapi.yaml \
.env.example \
deploy/.env.example \
deploy/helm/certctl/values.yaml \
deploy/helm/certctl/templates/ \
2>/dev/null \
| grep -v '_test.go' \
| grep -vE '^\s*[^:]+:[0-9]+:\s*(//|#)' \
| grep -v 'is no longer accepted' \
|| true)
if [ -n "$BAD" ]; then
echo "G-1 regression: \"jwt\" reappeared in an allowed-set surface:"
echo "$BAD"
echo ""
echo "Allowed surface for 'jwt' literals: comment lines, the"
echo "dedicated rejection branch in internal/config/config.go,"
echo "and connector packages (Google OAuth2, step-ca)."
echo "See docs/upgrade-to-v2-jwt-removal.md and"
echo "internal/config/config.go::ValidAuthTypes()."
exit 1
fi
- name: Forbidden api_key_hash JSON-shape regression guard (G-2)
# G-2 closed cat-s5-apikey_leak by tagging Agent.APIKeyHash
# `json:"-"` and adding a defense-in-depth Agent.MarshalJSON that
# zeroes the field on the marshal-time copy. This step grep-fails
# the build if `api_key_hash` reappears in any of the *additive*
# JSON-emitting surfaces: a Go struct json tag in internal/domain/,
# an OpenAPI Agent schema property, a TypeScript field declaration
# in web/src/, or an enum-list / discriminator in handler
# production code.
#
# Repository, migration, seed, service, integration-test, and
# unit-test files are exempt — those are server-internal use
# sites (the DB column stays, the in-memory struct field stays,
# the auth-lookup path stays). Comment lines are exempt so the
# G-2 closure rationale can stay in the source.
#
# See coverage-gap-audit-2026-04-24-v5/unified-audit.md
# cat-s5-apikey_leak for the closure rationale, or
# internal/domain/connector.go::Agent::MarshalJSON for the
# redaction enforcement.
run: |
set -e
# Scoped patterns that indicate api_key_hash being added back
# to a JSON-emitting surface. Each catches a regression shape
# that pre-G-2 actually shipped or that a future refactor
# could plausibly introduce:
# - Go struct tag: `json:"api_key_hash"`
# - Frontend interface: api_key_hash[?]: string
# - OpenAPI schema property: api_key_hash: (column-aligned)
# - YAML enum / array: - api_key_hash
BAD=$(grep -rnEH \
-e 'json:"api_key_hash[",]' \
-e '^\s*api_key_hash\??\s*:' \
-e '^\s*-\s*api_key_hash\s*$' \
internal/domain/ \
internal/api/ \
cmd/ \
api/openapi.yaml \
web/src/ \
2>/dev/null \
| grep -v '_test.go' \
| grep -vE '^\s*[^:]+:[0-9]+:\s*(//|#)' \
|| true)
if [ -n "$BAD" ]; then
echo "G-2 regression: api_key_hash reappeared in a JSON-emitting surface:"
echo "$BAD"
echo ""
echo "Allowed surface for api_key_hash literals: comment lines,"
echo "the database column (migrations/), the in-memory struct"
echo "field tagged \`json:\"-\"\`, and the repository / service"
echo "use sites. See internal/domain/connector.go::Agent and"
echo "coverage-gap-audit-2026-04-24-v5/unified-audit.md"
echo "cat-s5-apikey_leak for the closure rationale."
exit 1
fi
- name: Forbidden plaintext HEALTHCHECK regression guard (U-2)
# U-2 closed cat-u-healthcheck_protocol_mismatch by switching the
# published image's HEALTHCHECK from `curl -f http://localhost:
# 8443/health` (always failed against the HTTPS-only listener) to
# `curl -fsk https://localhost:8443/health`. This step grep-fails
# the build if any Dockerfile in the repo carries the pre-U-2
# plaintext shape — either explicitly (`http://localhost:8443/
# health` in a HEALTHCHECK) or via the looser pattern of any
# HEALTHCHECK that targets `http://` against the certctl server
# port.
#
# Comment lines and the docs/upgrade-to-tls.md:182 expected-to-
# fail invariant ("plaintext is gone, expect Connection refused")
# are intentionally exempt — we DO want the upgrade-doc string
# `http://localhost:8443/health` to remain there, since it
# documents what operators should test for to confirm plaintext
# is dead. The guardrail is scoped to Dockerfile* only, so docs
# are out of its reach.
#
# See coverage-gap-audit-2026-04-24-v5/unified-audit.md
# cat-u-healthcheck_protocol_mismatch for the closure rationale,
# or deploy/test/healthcheck_test.go for the binary-image
# contract the runtime test pins.
run: |
set -e
# Patterns that catch the actual regression shapes:
# - HEALTHCHECK directive carrying any http:// (even if the
# port differs, no plaintext probe should ship).
# - The exact pre-U-2 string for grep-friendliness.
BAD=$(grep -rnEH \
-e 'HEALTHCHECK.*http://' \
-e 'curl[^|&;]*-f[^|&;]*http://localhost:8443/health' \
Dockerfile Dockerfile.agent Dockerfile.* 2>/dev/null \
| grep -vE '^\s*[^:]+:[0-9]+:\s*#' \
|| true)
if [ -n "$BAD" ]; then
echo "U-2 regression: plaintext HEALTHCHECK reappeared in a Dockerfile:"
echo "$BAD"
echo ""
echo "Allowed: HTTPS HEALTHCHECK with -k (acceptable for"
echo "localhost-to-localhost), or non-HTTP probe shapes"
echo "(pgrep, /proc check). See Dockerfile / Dockerfile.agent"
echo "for the post-U-2 reference shape and"
echo "coverage-gap-audit-2026-04-24-v5/unified-audit.md"
echo "cat-u-healthcheck_protocol_mismatch for rationale."
exit 1
fi
- name: Race Detection
run: go test -race ./internal/service/... ./internal/api/handler/... ./internal/api/middleware/... ./internal/scheduler/... ./internal/connector/... ./internal/crypto/... ./internal/domain/... ./internal/validation/... ./internal/tlsprobe/... -count=1 -timeout 300s
- name: Go Test with Coverage
run: |
go test ./internal/service/... ./internal/api/handler/... ./internal/api/middleware/... ./internal/integration/... ./internal/connector/issuer/... ./internal/connector/target/... ./internal/connector/notifier/... ./internal/connector/discovery/... ./internal/crypto/... ./internal/mcp/... ./internal/cli/... ./internal/domain/... ./internal/validation/... ./internal/tlsprobe/... -count=1 -cover -coverprofile=coverage.out
- name: Check Coverage Thresholds
run: |
# Extract per-package coverage from test output
echo "=== Coverage Report ==="
go tool cover -func=coverage.out | tail -1
# Check service layer coverage (target: 60%+)
SERVICE_COV=$(go tool cover -func=coverage.out | grep 'internal/service' | awk '{print $NF}' | sed 's/%//' | awk '{sum+=$1; n++} END {if(n>0) printf "%.1f", sum/n; else print "0"}')
echo "Service layer coverage: ${SERVICE_COV}%"
# Check handler layer coverage (target: 60%+)
HANDLER_COV=$(go tool cover -func=coverage.out | grep 'internal/api/handler' | awk '{print $NF}' | sed 's/%//' | awk '{sum+=$1; n++} END {if(n>0) printf "%.1f", sum/n; else print "0"}')
echo "Handler layer coverage: ${HANDLER_COV}%"
# Check domain layer coverage (target: 40%+)
DOMAIN_COV=$(go tool cover -func=coverage.out | grep 'internal/domain' | awk '{print $NF}' | sed 's/%//' | awk '{sum+=$1; n++} END {if(n>0) printf "%.1f", sum/n; else print "0"}')
echo "Domain layer coverage: ${DOMAIN_COV}%"
# Check middleware layer coverage (target: 50%+)
MIDDLEWARE_COV=$(go tool cover -func=coverage.out | grep 'internal/api/middleware' | awk '{print $NF}' | sed 's/%//' | awk '{sum+=$1; n++} END {if(n>0) printf "%.1f", sum/n; else print "0"}')
echo "Middleware layer coverage: ${MIDDLEWARE_COV}%"
# Check crypto package coverage (target: 85%+)
# M-8 rationale: encryption primitives are a security-critical gate.
# v2 format, key-derivation, fallback, and fail-closed sentinel paths
# all need exhaustive coverage to avoid silent regressions (CWE-916 / CWE-329).
CRYPTO_COV=$(go tool cover -func=coverage.out | grep 'internal/crypto' | awk '{print $NF}' | sed 's/%//' | awk '{sum+=$1; n++} END {if(n>0) printf "%.1f", sum/n; else print "0"}')
echo "Crypto package coverage: ${CRYPTO_COV}%"
# Fail if thresholds not met
if [ "$(echo "$SERVICE_COV < 55" | bc -l)" -eq 1 ]; then
echo "::error::Service layer coverage ${SERVICE_COV}% is below 55% threshold"
exit 1
fi
if [ "$(echo "$HANDLER_COV < 60" | bc -l)" -eq 1 ]; then
echo "::error::Handler layer coverage ${HANDLER_COV}% is below 60% threshold"
exit 1
fi
if [ "$(echo "$DOMAIN_COV < 40" | bc -l)" -eq 1 ]; then
echo "::error::Domain layer coverage ${DOMAIN_COV}% is below 40% threshold"
exit 1
fi
if [ "$(echo "$MIDDLEWARE_COV < 30" | bc -l)" -eq 1 ]; then
echo "::error::Middleware layer coverage ${MIDDLEWARE_COV}% is below 30% threshold"
exit 1
fi
if [ "$(echo "$CRYPTO_COV < 85" | bc -l)" -eq 1 ]; then
echo "::error::Crypto package coverage ${CRYPTO_COV}% is below 85% threshold"
exit 1
fi
echo "Coverage thresholds passed!"
- name: Upload Coverage Report
uses: actions/upload-artifact@v4
with:
name: go-coverage
path: coverage.out
retention-days: 30
frontend-build:
name: Frontend Build
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Node.js
uses: actions/setup-node@v4
with:
node-version: '22'
- name: Install Dependencies
working-directory: web
run: npm ci
- name: TypeScript Check
working-directory: web
run: npx tsc --noEmit
- name: Run Frontend Tests
working-directory: web
run: npx vitest run
- name: Build Frontend
working-directory: web
run: npx vite build
helm-lint:
name: Helm Chart Validation
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install Helm
uses: azure/setup-helm@v4
with:
version: '3.13.0'
# HTTPS-Everywhere (v2.0.47): the chart fails render when no TLS source is
# configured. Every lint/template invocation below must pick exactly one
# provisioning mode — see deploy/helm/certctl/templates/_helpers.tpl
# (certctl.tls.required) and docs/tls.md.
- name: Lint Helm Chart
run: |
helm lint deploy/helm/certctl/ \
--set server.tls.existingSecret=certctl-tls-ci
- name: Template Helm Chart (existingSecret mode)
run: |
helm template certctl deploy/helm/certctl/ \
--set server.tls.existingSecret=certctl-tls-ci \
> /dev/null
- name: Template Helm Chart (cert-manager mode)
run: |
helm template certctl deploy/helm/certctl/ \
--set server.tls.certManager.enabled=true \
--set server.tls.certManager.issuerRef.name=letsencrypt-prod \
> /dev/null
- name: Template Helm Chart (guard fails without TLS)
run: |
# Inverse test: the chart MUST refuse to render when no TLS source is
# configured. If this ever renders successfully, the fail-loud guard
# in certctl.tls.required has regressed.
if helm template certctl deploy/helm/certctl/ > /dev/null 2>&1; then
echo "::error::Helm chart rendered without a TLS source — fail-loud guard regressed"
exit 1
fi