infra-kubernetes-cluster-health

What problem does it solve?

It helps you identify why a Kubernetes cluster or workloads are unhealthy by pinpointing node pressure, pod failure causes, and control-plane issues that lead to outages or degraded scheduling.

Core Features & Use Cases

• Node health assessment: Evaluate node readiness and specific pressure conditions (MemoryPressure, DiskPressure, PIDPressure, NetworkUnavailable) and take the right remediation steps (drain/cordon/eviction).
• Pod failure diagnosis: Quickly triage CrashLoopBackOff, OOMKilled, Pending, and Evicted pods using targeted kubectl commands and event/log evidence.
• Control plane verification: Check API server responsiveness and etcd health, plus scheduler/controller-manager component health.
• Scheduling and quota troubleshooting: Diagnose quota/limits problems, taints/tolerations mismatches, affinity/nodeSelector issues, and insufficient allocatable resources.
• Operational runbooks: Provide a repeatable “cluster health summary” workflow and alert rules to catch issues early.

Quick Start

Use this skill to troubleshoot a production incident by asking: "Assess my Kubernetes cluster health and diagnose why my pods are CrashLooping and some nodes are under MemoryPressure."