How to diagnose a crashing container in a pod (CrashLoopBackOff)?

To diagnose a container in CrashLoopBackOff state, follow these steps:

1. Get the list of pods and their statuses:

kubectl get pods

2. Check the logs of the crashing container:

kubectl logs <pod-name> --previous

The --previous flag allows you to see the logs of the previous container run if it is restarting.

3. Describe the pod for errors and events:

kubectl describe pod <pod-name>

Pay attention to the Events section — it may contain reasons for restarts.

4. If the container crashes due to an application error, logs will help identify the cause (e.g., exception, configuration error).

5. If the container crashes due to resource issues (e.g., OOMKilled), this will be visible in the pod description.

6. If necessary, connect to the container (if it starts but crashes later) for additional diagnostics:

kubectl exec -it <pod-name> -- /bin/sh

7. Check readiness and liveness probes, as they might be misconfigured and cause restarts.

Thus, the main diagnostics involve viewing logs, describing the pod, and analyzing events.