infra-kubernetes-cluster-health

Diagnose Kubernetes pod failures by gathering evidence with kubectl commands like describe, get, and logs to identify CrashLoopBackOff or OOMKilled causes. You can then apply targeted checks for scheduling constraints and resource quotas to pinpoint the root cause.

Kubernetes node pressure conditions like MemoryPressure, DiskPressure, or PIDPressure occur when system resources are exhausted. Assess node readiness and specific pressure states using kubectl evidence to determine the right remediation steps such as drain, cordon, or eviction.

Verify Kubernetes control plane health by checking API server responsiveness and etcd status using kubectl. This process evaluates scheduler and controller-manager component health to determine if control-plane unresponsiveness is causing workload outages.

Yes, you can troubleshoot Kubernetes scheduling failures by checking taints and tolerations mismatches, affinity issues, and insufficient allocatable resources. Diagnose quota and limits problems using targeted kubectl commands to resolve degraded scheduling.

The best way to triage Evicted pods is using a repeatable cluster health summary workflow with kubectl-based evidence gathering from events and metrics. This approach identifies node pressure and provides actionable remediation guidance to prevent workload outages.