메뉴
Debug Kubernetes pods, nodes, and workloads. Use when pods are failing, containers crash, nodes are unhealthy, or users mention debugging, troubleshooting, or diagnosing Kubernetes issues.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Configure Kubernetes autoscaling with HPA, VPA, and KEDA. Use for horizontal/vertical pod autoscaling, event-driven scaling, and capacity management.
Kubernetes backup and restore with Velero. Use when creating backups, restoring applications, managing disaster recovery, or migrating workloads between clusters.
Browser automation for Kubernetes dashboards and web UIs. Use when interacting with Kubernetes Dashboard, Grafana, ArgoCD UI, or other web interfaces. Requires MCP_BROWSER_ENABLED=true.
Cluster API lifecycle management for provisioning, scaling, and upgrading Kubernetes clusters. Use when managing cluster infrastructure or multi-cluster operations.
Kubernetes certificate management with cert-manager. Use when managing TLS certificates, configuring issuers, or troubleshooting certificate issues.
Cilium and Hubble network observability for Kubernetes. Use when managing network policies, observing traffic flows, or troubleshooting connectivity with eBPF-based networking.
| name | k8s-troubleshoot |
| description | Debug Kubernetes pods, nodes, and workloads. Use when pods are failing, containers crash, nodes are unhealthy, or users mention debugging, troubleshooting, or diagnosing Kubernetes issues. |
| license | Apache-2.0 |
| metadata | {"author":"rohitg00","version":"1.0.0","tools":15,"category":"observability"} |
Expert debugging and diagnostics for Kubernetes clusters using kubectl-mcp-server tools.
Use this skill when:
| Priority | Rule | Impact | Tools |
|---|---|---|---|
| 1 | Check pod status first | CRITICAL | get_pods, describe_pod |
| 2 | View recent events | CRITICAL | get_events |
| 3 | Inspect logs (including previous) | HIGH | get_pod_logs |
| 4 | Check resource metrics | HIGH | get_pod_metrics |
| 5 | Verify endpoints | MEDIUM | get_endpoints |
| 6 | Review network policies | MEDIUM | get_network_policies |
| 7 | Examine node status | LOW | get_nodes, describe_node |
| Symptom | First Tool | Next Steps |
|---|---|---|
| Pod Pending | describe_pod | Check events, node capacity, resource requests |
| CrashLoopBackOff | get_pod_logs(previous=True) | Check exit code, resources, liveness probes |
| ImagePullBackOff | describe_pod | Verify image name, registry auth, network |
| OOMKilled | get_pod_metrics | Increase memory limits, check for memory leaks |
| ContainerCreating | describe_pod | Check PVC binding, secrets, configmaps |
| Terminating (stuck) | describe_pod | Check finalizers, PDBs, preStop hooks |
1. get_pods(namespace, label_selector) - Get pod status
2. describe_pod(name, namespace) - See events and conditions
3. get_events(namespace, field_selector="involvedObject.name=<pod>") - Check events
4. get_pod_logs(name, namespace, previous=True) - For crash loops
| State | Likely Cause | Tools to Use |
|---|---|---|
| Pending | Scheduling issues | describe_pod, get_nodes, get_events |
| ImagePullBackOff | Registry/auth | describe_pod, check image name |
| CrashLoopBackOff | App crash | get_pod_logs(previous=True) |
| OOMKilled | Memory limit | get_pod_metrics, adjust limits |
| ContainerCreating | Volume/network | describe_pod, get_pvc |
1. get_nodes() - List nodes and status
2. describe_node(name) - See conditions and capacity
3. Check: Ready, MemoryPressure, DiskPressure, PIDPressure
4. node_logs_tool(name, "kubelet") - Kubelet logs
1. get_pod_logs(name, namespace, previous=True) - See why it crashed
2. describe_pod(name, namespace) - Check resource limits, probes
3. get_pod_metrics(name, namespace) - Memory/CPU at crash time
4. If OOM: compare requests/limits to actual usage
5. If app error: check logs for stack trace
1. get_services(namespace) - Verify service exists
2. get_endpoints(namespace) - Check endpoint backends
3. If empty endpoints: pods don't match selector
4. get_network_policies(namespace) - Check traffic rules
5. For Cilium: cilium_endpoints_list_tool(), hubble_flows_query_tool()
1. get_pvc(namespace) - Check PVC status
2. describe_pvc(name, namespace) - See binding issues
3. get_storage_classes() - Verify provisioner exists
4. If Pending: check storage class, access modes
1. kubectl_exec(pod, namespace, "nslookup kubernetes.default") - Test DNS
2. If fails: check coredns pods in kube-system
3. get_pods(namespace="kube-system", label_selector="k8s-app=kube-dns")
4. get_pod_logs(name="coredns-*", namespace="kube-system")
All tools support context parameter for targeting different clusters:
get_pods(namespace="kube-system", context="production-cluster")
get_events(namespace="default", context="staging-cluster")
describe_pod(name="myapp-xyz", namespace="prod", context="prod-east")
For comprehensive diagnostics, run the bundled scripts:
See references/DECISION-TREE.md for visual troubleshooting flowcharts.
See references/COMMON-ERRORS.md for error message explanations and fixes.
get_pods, describe_pod, get_pod_logs, get_pod_metricsget_events, get_nodes, describe_nodeget_resource_usage, compare_namespacescilium_endpoints_list_tool, hubble_flows_query_toolistio_proxy_status_tool, istio_analyze_tool