Lab 6: Deploy & Scale¶

Field	Value
Tier	2 — Kubernetes Core
Estimated Time	45 minutes
Prerequisites	k3s cluster, kubectl
Auto-Grade	Yes

Scenario¶

Your company is migrating a three-tier web application from VMs to Kubernetes. The application has a React frontend served by nginx, a Python Flask API server, and a PostgreSQL database. The previous ops team started writing Kubernetes manifests but left the project half-done. The deployments exist but are missing resource limits, the frontend has only one replica (it should have three), and the API server keeps getting OOM-killed because nobody set memory limits.

The product manager wants this deployed and stable before the Monday demo. You need to complete the manifests, deploy all three tiers, configure proper resource requests and limits, set up a Horizontal Pod Autoscaler for the API, and verify the entire stack is healthy.

Objectives¶

Deploy frontend with 3 replicas, each with resource requests (cpu: 100m, memory: 128Mi) and limits (cpu: 200m, memory: 256Mi)
Deploy API server with 2 replicas, resource limits, and readiness probe on /health
Deploy PostgreSQL as a single-replica Deployment with a PVC for data
Create ClusterIP services for all three tiers
Configure HPA for the API (min 2, max 8, target CPU 70%)
All pods are Running and Ready within 5 minutes
Frontend pods can reach the API service via DNS

Setup¶

./setup.sh

Creates a namespace lab-deploy with incomplete manifests applied.

Hints¶

Hint 1: Resource requests and limits

Add under `spec.containers[].resources`:

resources:
  requests:
    cpu: "100m"
    memory: "128Mi"
  limits:
    cpu: "200m"
    memory: "256Mi"

Hint 2: Readiness probes

readinessProbe:
  httpGet:
    path: /health
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 10

Hint 3: HPA configuration

`kubectl autoscale deployment api --min=2 --max=8 --cpu-percent=70 -n lab-deploy`

Hint 4: Service DNS

Services are reachable at `..svc.cluster.local`. Use `kubectl exec` into a frontend pod and `curl http://api.lab-deploy.svc.cluster.local:8080/health`.

Grading¶

./grade.sh

Solution¶

See the solution/ directory for complete manifests.