Health Endpoints

The server exposes two endpoints for container orchestrators and external health checks.

/health

GET /health
200 OK
{"status": "ok"}

Always returns 200 if the HTTP listener is responding. Suitable for a liveness probe — its only signal is “the process is alive enough to answer HTTP”. A failing pipeline or stuck rule engine will not flip this to a non-200 response.

/ready

GET /ready
200 OK   {"status": "ready"}     — supervisor has spawned the pipelines
503      {"status": "not_ready"} — startup not complete

Returns 200 once the supervisor has finished spawning all configured pipelines. Suitable for a readiness probe.

Note

/ready flips to 200 once pipelines have been spawned, not necessarily connected. A misconfigured database (wrong host, bad credentials) can leave a pipeline in a reconnect loop while /ready still returns 200. For per-database connection state, scrape pgsense_pipeline_connected{database="..."} from /metrics and alert on it directly.

Kubernetes probes

livenessProbe:
  httpGet:
    path: /health
    port: 9090
  periodSeconds: 10
  failureThreshold: 3
readinessProbe:
  httpGet:
    path: /ready
    port: 9090
  periodSeconds: 10
  failureThreshold: 3
  initialDelaySeconds: 5

initialDelaySeconds on the readiness probe gives the supervisor time to spawn pipelines before the orchestrator decides the pod is unready.

Docker / Compose

services:
  pgsense:
    image: ghcr.io/mrehbr/pgsense-rs:latest
    healthcheck:
      test: ["CMD", "curl", "-fsS", "http://localhost:9090/ready"]
      interval: 10s
      timeout: 3s
      retries: 3
      start_period: 10s