kubernetes

Kubernetes Best Practices

Workload configuration

Pod resource management

• Always set resource requests AND limits
• Use resources.requests for scheduling, resources.limits for throttling
• Start conservative and tune based on metrics

resources:
  requests:
    cpu: 100m
    memory: 128Mi
  limits:
    cpu: 500m
    memory: 512Mi

Health checks

• livenessProbe — restart if unhealthy (use for deadlock detection)
• readinessProbe — remove from service if not ready (use for startup/dependency checks)
• startupProbe — delay liveness checks during startup (use for slow-starting apps)

Pod disruption budgets

• Always define PDBs for production workloads
• minAvailable: 1 or maxUnavailable: 1 for small deployments

Troubleshooting

Pod not starting

1. kubectl describe pod <name> — check Events section
2. kubectl logs <name> --previous — check crash logs
3. Common causes: image pull errors, resource limits, missing secrets

Service not reachable

1. kubectl get endpoints <service> — verify endpoints exist
2. kubectl get pods -l <selector> — check pod readiness
3. kubectl exec -it <pod> -- curl localhost:<port>/health — test from inside

OOMKilled

1. Check kubectl describe pod for last termination reason
2. Increase memory limits
3. Profile the application for memory leaks

Security

• Use NetworkPolicies to restrict pod-to-pod traffic
• Never run containers as root — use securityContext.runAsNonRoot: true
• Use ServiceAccounts with minimal RBAC permissions
• Scan images for vulnerabilities before deploying