Case Studies¶
Every earlier section teaches one concept at a time. These case studies put several together the way a real production task actually requires — a realistic scenario, real manifests, real kubectl commands, and a failure path with its fix, not just the happy path.
Full Case Studies¶
- Rolling Deployment with Zero Downtime — tuned readiness probes and
maxSurge/maxUnavailable, verified with a live request-loss counter - Blue-Green and Canary Releases — a Service-selector cutover and a weighted-Ingress canary, both with a rollback path
- Multi-Tenant Namespace Setup — onboarding a new team with
ResourceQuota,LimitRange, RBAC, and default-denyNetworkPolicy - Autoscaling Under Load — HPA on CPU and a custom metric during a real traffic spike, tuned to stop flapping
- Secrets Rotation — rotating a database credential with no pod ever holding the stale value and no downtime
- Debugging a CrashLoopBackOff Incident — a bad deploy, a real incident narrative, and the
describe/logs --previous/eventstrail that finds the root cause
Next¶
Continue to Troubleshooting.