serverpod-health-checks

Configure your Serverpod instance with HealthConfig to expose Kubernetes-ready health endpoints, using /livez for liveness, /readyz for readiness, and /startupz for startup probes on port 8080.

In Serverpod health checks, the liveness endpoint confirms the process is alive without checking dependencies, while the readiness endpoint reports whether dependencies like DB and Redis are healthy, returning 503 when they are not.

Yes, you can implement custom HealthIndicator classes to check external services and register them via HealthConfig, alongside built-in indicators like DatabaseHealthIndicator and RedisHealthIndicator.

Yes, Serverpod health checks support authenticated RFC JSON health responses for detailed monitoring, as well as unauthenticated HTTP status-only responses returning 200 or 503 for Kubernetes probes.

You can configure per-indicator timeouts and caching through HealthConfig when registering built-in and additional custom HealthIndicator implementations for your Serverpod backend service.

Use the /startupz endpoint to indicate the server has finished initializing, which is necessary when your Serverpod backend requires significant startup time before Kubernetes should begin liveness and readiness probing.