Module 3 Exercises: Working with Objects

In Module 3, we moved from architecture to action. You now understand how to define the core building blocks of any Kubernetes application. These exercises will guide you through the process of creating a simple, stateful application stack.

Exercise 1: The "Simple AI" Pod

Create a YAML manifest for a Pod named ai-hello-world.

Image: Use python:3.11-slim or your own test image.
Environment Variable: Add a variable MODEL_TYPE with value experimental.
Resources: Set a CPU request of 100m and a memory request of 128Mi.
Verification: Once applied, run kubectl get pod ai-hello-world -o yaml and find the uid of the pod.

Exercise 2: Scaling the AI Engine

Convert your single Pod into a Deployment.

Deployment Name: ai-engine-deployment
Replicas: 3
Update Strategy: RollingUpdate with 0 unavailable pods.
Labels: Use app: ai-engine for both the selector and the pod template.
Scaling: Use kubectl scale to increase the replicas to 5, then back to 3.

Exercise 3: Exposing the Internal Service

Create a Service to allow other pods to talk to your ai-engine-deployment.

Name: ai-engine-service
Type: ClusterIP
Ports: Map port 80 to the target port 8000.
Verification: Run kubectl get endpoints ai-engine-service. Do you see multiple IP addresses? Why?

Exercise 4: Persistent Data Storage

Create a PersistentVolumeClaim for a database.

Name: db-storage-pvc
Size: 1Gi
Access Mode: ReadWriteOnce
Integration: Add this PVC as a volume to a new Deployment named db-deployment. Mount it at /var/lib/mysql or similar.

Solutions (Self-Check)

Exercise 1 Strategy:

Your YAML should look like this in the spec section:

containers:
- name: hello
  image: python:3.11-slim
  env:
  - name: MODEL_TYPE
    value: experimental
  resources:
    requests:
      cpu: 100m
      memory: 128Mi

Exercise 2 Hint:

The kubectl scale commands are:

kubectl scale deployment ai-engine-deployment --replicas=5
kubectl scale deployment ai-engine-deployment --replicas=3

Exercise 3 Solution:

Yes, you should see multiple IPs in the Endpoints list! One IP for each "Ready" pod managed by the deployment. The Service uses these to load balance traffic.

Exercise 4 Hint:

Inside the spec.template.spec of your Deployment:

volumes:
- name: db-data
  persistentVolumeClaim:
    claimName: db-storage-pvc
containers:
- name: db-container
  image: mysql
  volumeMounts:
  - name: db-data
    mountPath: /var/lib/mysql

Summary of Module 3

Congratulations! You are no longer just a theorist; you are a Practitioner.

You can define Pods and understand their lifecycle.
You can manage Deployments and update them with zero downtime.
You can build stable gateways using Services.
You can manage Secrets and ConfigMaps.
You can give your apps a permanent memory with PVCs.

In Module 4: Deploying Applications, we will learn how to automate this process and manage complex, multi-container applications in the real world.