vincent119/k8s-debug

K8s Debugging Workflow

Systematic approach to diagnose and fix Kubernetes application issues.

1. Check Pod Status

Identify pods in error states (CrashLoopBackOff, ImagePullBackOff, etc.):

kubectl get pods -o wide
kubectl describe pod <POD_NAME>

2. View Logs

Check container logs for errors and stack traces:

# Current logs
kubectl logs <POD_NAME> [-c <CONTAINER>]

# Previous crash logs
kubectl logs <POD_NAME> --previous

3. Check Events

Review cluster events for scheduling/mounting/health check failures:

kubectl get events --sort-by=.lastTimestamp

4. Exec into Container

Debug filesystem, environment, or network issues:

kubectl exec -it <POD_NAME> -- sh

# Inside container:
env                          # Check environment variables
curl localhost:8080          # Test HTTP endpoints
nc -zv <HOST> <PORT>         # Test network connectivity

5. Port Forward

Forward pod port to local for testing:

kubectl port-forward <POD_NAME> 8080:8080

6. Monitor Resources

Check for OOM or CPU throttling:

kubectl top pod <POD_NAME>

Common Issues

| Issue | Command | Solution | |-------|---------|----------| | CrashLoopBackOff | logs --previous | Check startup errors | | ImagePullBackOff | describe pod | Verify image name/credentials | | Pending | get events | Check resource limits/node capacity | | OOMKilled | top pod | Increase memory limits |