SKILL.md
Kubernetes Diagnostics
Analyze cluster health and compare resources using kubectl-mcp-server's diagnostic tools.
When to Apply
Use this skill when:
- User mentions: "metrics", "health check", "compare", "analysis", "capacity"
- Operations: cluster health assessment, environment comparison, resource analysis
- Keywords: "how much", "usage", "difference between", "capacity planning"
Priority Rules
| Priority | Rule | Impact | Tools |
|---|---|---|---|
| 1 | Check metrics-server before using metrics | CRITICAL | getresourcemetrics |
| 2 | Run health check before deployments | HIGH | clusterhealthcheck |
| 3 | Compare staging vs prod before release | MEDIUM | compare_namespaces |
| 4 | Document baseline metrics | LOW | getnodessummary |
Quick Reference
| Task | Tool | Example |
|---|---|---|
| Cluster health | clusterhealthcheck |
clusterhealthcheck() |
| Pod metrics | getresourcemetrics |
getresourcemetrics(namespace) |
| Node summary | getnodessummary |
getnodessummary() |
| Compare envs | compare_namespaces |
compare_namespaces(ns1, ns2, type) |
| List CRDs | list_crds |
list_crds() |
Resource Metrics
get_resource_metrics(namespace="default")
get_node_metrics()
get_top_pods(namespace="default", sort_by="cpu")
get_top_pods(namespace="default", sort_by="memory")Cluster Health Check
cluster_health_check()
get_cluster_info()Compare Environments
compare_namespaces(
namespace1="staging",
namespace2="production",
resource_type="deployment"
)
compare_namespaces(
namespace1="default",
namespace2="default",
resource_type="deployment",
context1="staging-cluster",
context2="prod-cluster"
)API Discovery
get_api_versions()
check_crd_exists(crd_name="certificates.cert-manager.io")
list_crds()Resource Analysis
get_nodes_summary()
kubeconfig_view()
list_contexts_tool()Diagnostic Workflows
Cluster Overview
cluster_health_check()
get_nodes_summary()
get_events(namespace="")
list_crds()Pre-deployment Check
get_resource_metrics(namespace="production")
get_nodes_summary()
compare_namespaces(namespace1="staging", namespace2="prod", resource_type="deployment")Post-incident Analysis
get_events(namespace)
get_pod_logs(name, namespace, previous=True)
get_resource_metrics(namespace)
describe_node(name)Related Skills
- [k8s-troubleshoot](../k8s-troubleshoot/SKILL.md) - Debug issues
- [k8s-cost](../k8s-cost/SKILL.md) - Cost analysis
- [k8s-incident](../k8s-incident/SKILL.md) - Incident response