init

.opencode/skills/devops/references/kubernetes-troubleshooting.md

@@ -0,0 +1,49 @@
+# Kubernetes Troubleshooting
+
+## Debugging Workflow
+
+```bash
+# 1. Overview
+kubectl get pods -o wide
+kubectl get events -n <namespace> --sort-by='.lastTimestamp'
+
+# 2. Details
+kubectl describe pod <pod-name>
+
+# 3. Logs
+kubectl logs <pod-name>
+kubectl logs <pod-name> --previous  # Crashed instance
+kubectl logs <pod-name> -c <container>
+```
+
+## Common Pod States
+
+| State | Cause | Solution |
+|-------|-------|----------|
+| Pending | No node resources | Check node capacity |
+| ContainerCreating | Image pulling | Check image URI |
+| CrashLoopBackOff | Container exits | Check logs, health checks |
+| ImagePullBackOff | Failed image pull | Verify credentials |
+| OOMKilled (137) | Out of memory | Increase memory limit |
+
+## Service & Network
+
+```bash
+kubectl exec -it <pod-name> -- nslookup kubernetes.default
+kubectl exec -it <pod-name> -- curl http://myservice:8080
+kubectl get endpoints <service-name>
+kubectl port-forward service/myservice 8080:8080
+kubectl get networkpolicies -A
+```
+
+## Quick Fixes
+
+| Problem | Command |
+|---------|---------|
+| Pod stuck | `kubectl delete pod <name> --grace-period=0 --force` |
+| High CPU | `kubectl top pods -A --sort-by=cpu` |
+| High memory | `kubectl top pods -A --sort-by=memory` |
+| Restart | `kubectl rollout restart deployment/<name>` |
+| Rollback | `kubectl rollout undo deployment/<name>` |
+
+See `kubernetes-troubleshooting-advanced.md` for node issues, HPA, anti-patterns.