Kubernetes "Readiness probe failed" - pod not serving traffic in CI

A readiness probe gates whether a pod receives traffic. While it fails, the pod stays out of the Service endpoints and the Deployment never reports enough Ready replicas - so a rollout status in CI hangs even though the container is "Running".

What this error means

kubectl get pods shows 0/1 Ready though Running; describe shows Readiness probe failed; kubectl get endpoints <svc> is empty. The deploy step waits and eventually times out.

kubectl

NAME                  READY   STATUS    RESTARTS   AGE
api-6b4c9d-xyz        0/1     Running   0          2m
Warning  Unhealthy  kubelet  Readiness probe failed: HTTP probe failed with statuscode: 503

Common causes

App not finished warming up

The readiness check fires before caches/connections are ready, so during a slow rollout the pod stays NotReady longer than the probe/deadline allows.

Wrong readiness endpoint or a real dependency outage

The probe path/port is wrong, or a downstream the readiness check depends on (DB, queue) is unavailable.

How to fix it

Check readiness state and endpoints

Confirm whether it is timing (slow warmup) or a genuine failure.

Terminal

kubectl get pod <pod> -o jsonpath='{.status.conditions}'
kubectl get endpoints <svc>
kubectl describe pod <pod> | grep -i readiness

Fix the probe or the dependency

Make the readiness endpoint reflect real serving capability, not just process-up.
Loosen initialDelay/period/failureThreshold if warmup legitimately takes longer.
If a dependency is down, the readiness probe is doing its job - fix the dependency.

How to prevent it

Keep readiness checks fast and tied to actual serving readiness.
Tune timings to measured warmup, and add a startupProbe for long boots.
Set a sane progressDeadlineSeconds so a stuck rollout fails clearly.