Skip to content

feat: smoke tests use /health endpoints#18

Closed
DarlingtonDeveloper wants to merge 5 commits into
mainfrom
feat/health-endpoint-smoke-tests
Closed

feat: smoke tests use /health endpoints#18
DarlingtonDeveloper wants to merge 5 commits into
mainfrom
feat/health-endpoint-smoke-tests

Conversation

@DarlingtonDeveloper
Copy link
Copy Markdown
Collaborator

Updates smoke test to use standardised /health endpoints with JSON status parsing.

Changes:

  • Services with published ports (Alexandria, Dispatch, Chronicle, Dredd) now use /health endpoints
  • Parses JSON response to detect degraded state (e.g. DB down but service up)
  • Falls back to legacy endpoints if /health unavailable
  • Overlay-only services (PromptForge, Slack-gateway) checked via docker service ls replica count
  • NATS (Hermes) keeps existing healthz check
  • Slack alert format preserved, with optional degraded services block

Service was renamed but deploy configs still referenced the old name.
Updates service name, image, command, vault agent ID, and health check URL.
Dispatch /health endpoint returns 404 on current image (pre-PR #11).
Added /api/v1/backlog as fallback URL with X-Agent-ID header so the
smoke test passes until the image is rebuilt with the /health endpoint.
Instead of trying localhost for overlay-only services (Dispatch,
Chronicle, PromptForge, Slack-gateway), use docker exec against a
running container on the warren_agents network.

- Host checks: Alexandria (:8500), Dredd (:8750), NATS (:8222), OpenClaw (:18789)
- Overlay checks via probe: Dispatch, Chronicle, PromptForge, Slack-gateway
- Auto-discovers probe container (wget/curl/python3)
- Falls back to replica count if no probe available
- Slack alerting format unchanged
@DarlingtonDeveloper DarlingtonDeveloper deleted the feat/health-endpoint-smoke-tests branch February 16, 2026 15:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant