health-check-design

Liveness probes check process health to trigger restarts, while readiness probes check traffic-readiness and dependency health to route traffic. This Skill implements both for Go services to prevent restart-looping and ensure reliable traffic routing.

Implement graceful shutdown in Go by applying the not-ready-on-shutdown pattern to fail readiness probes. This Skill designs the handler and aligns termination grace periods to ensure Kubernetes performs rolling updates without dropping active connections.

Your pod restarts because the liveness probe is incorrectly checking dependency health instead of process health. This Skill separates liveness and readiness checks so transient dependency blips only remove the pod from service, avoiding self-inflicted restart loops.

Configure a startup probe to delay liveness checks until Go service initialization completes. This Skill designs startup, liveness, and readiness probe configurations to manage the full service lifecycle and prevent premature restarts during boot.

No, this Skill implements dependency-free liveness checks in Go. It designs timeout-bounded readiness probes that monitor dependency health without requiring external libraries.

The best way is distinguishing liveness, readiness, and startup probes to match the Go service lifecycle. This Skill designs production-ready health endpoints and probe configurations for zero-downtime deployments and reliable traffic routing.