Wrap-up

Hub › Kubernetes › 07

What you built

You took the beginner web Deployment from "it runs" to "it runs well":

Page	Change	Effect
02 ConfigMap	GREETING in web-config	Config without image rebuild
03 Probes	readinessProbe + livenessProbe	Traffic routing + auto-restart
04 Resources	requests + limits	Scheduler accuracy + OOM protection
05 Rolling update	nginx:1.27 → nginx:1.27.1	Zero-downtime image change
06 Scaling	1 → 4 → 2 replicas	Redundancy + load balancing

Each addition is independent. In a real cluster you'd apply all of them together from the start — this tier walked through them one at a time to make each concept visible.

Tear down

When you're done experimenting, turn off the local cluster. This is the same teardown as the beginner tier:

kubectl delete -f service.yaml -f deployment.yaml -f configmap.yaml
kind delete cluster --name learn

Confirm the cluster is gone:

kubectl get nodes   # error / no cluster — expected

Next steps

• Explore Horizontal Pod Autoscaler (kubectl autoscale) to scale based on CPU load rather than a fixed replica count.
• Try Ingress to expose multiple Services through a single load-balancer instead of port-forward.
• Look at Namespaces to isolate workloads on the same cluster.