Readiness probe failed

Kubernetes readiness probe connection refused

The kubelet reached the Pod network but no process accepted the readiness probe connection at the configured address and port.

kubelet probe workerHealth checkshigh

Exact message

Readiness probe failed: dial tcp: connect: connection refused

Also searched as: readiness probe failed · connection refused

What it means

The container can remain running while it is removed from Service endpoints. The application may not be listening yet, may listen on a different interface or port, or may have stopped accepting connections.

Typical trigger conditions

Wrong probe port or path
Application binds only to loopback
Startup takes longer than probe timing allows
Process is running but its listener failed

Evidence before intervention

First diagnostic checks

Compare probe with container ports

Inspect the deployed readinessProbe, named-port resolution and application listener configuration.

Read events and logs together

Align probe timestamps with application startup and shutdown messages.

kubectl describe pod <pod> -n <namespace>; kubectl logs <pod> -n <namespace> -c <container>

Test from the Pod network namespace

Use an approved debug container or equivalent to confirm the listener and response independently of the probe.

Version and platform notes

Probe-level terminationGracePeriodSeconds and gRPC probes require suitable Kubernetes versions; verify feature support.

Engineering signal

Follow Kubernetes failures and fixes

One useful weekly email with new error references, tools, integration notes and production lessons. No daily noise.