Kubernetes CrashLoopBackOff
A container repeatedly starts and exits, so the kubelet applies an increasing delay before trying again.
Back-off restarting failed containerAlso searched as: CrashLoopBackOff · BackOff
What it means
CrashLoopBackOff is a waiting reason, not a root cause. The previous termination state, exit code, events, logs and probe results identify why the process cannot remain healthy.
Typical trigger conditions
- Application process exits with an error
- Liveness probe repeatedly kills the container
- Missing configuration or secret
- Resource limit or dependency failure during startup
First diagnostic checks
Describe the Pod
Read current and last termination state plus ordered events.
kubectl describe pod <pod> -n <namespace>
Read previous container logs
The current container may be too new to contain the crash evidence.
kubectl logs <pod> -n <namespace> -c <container> --previous
Inspect exit reason and code
Distinguish application failure, signal termination and OOMKilled before changing probes or resources.
Version and platform notes
Restart backoff timing can be affected by kubelet feature gates in newer Kubernetes releases; always inspect cluster-version documentation.
Follow Kubernetes failures and fixes
One useful weekly email with new error references, tools, integration notes and production lessons. No daily noise.