Build and operate a service

Published on November 7, 2026

This lab follows one service from source code to a running Kubernetes Deployment. It uses a local kind cluster so you can inspect each step without a cloud account. Expect to finish with a reachable / endpoint and a passing /healthz readiness probe.

You need: Python 3, Docker, kind, kubectl, and a terminal. Run the commands from a new empty directory. The examples use a Unix-like shell.

Build a local service

Create app.py:


from http.server import BaseHTTPRequestHandler, HTTPServer
 
 
class Handler(BaseHTTPRequestHandler):
    def do_GET(self):
        if self.path == "/healthz":
            body = b"ok\n"
        elif self.path == "/":
            body = b"hello from the service\n"
        else:
            self.send_error(404)
            return
 
        self.send_response(200)
        self.send_header("Content-Type", "text/plain; charset=utf-8")
        self.send_header("Content-Length", str(len(body)))
        self.end_headers()
        self.wfile.write(body)
 
 
HTTPServer(("0.0.0.0", 8080), Handler).serve_forever()

Run python3 app.py in one terminal and curl -fsS http://localhost:8080/healthz in another. The expected response is ok. Stop the server before the next step.

Containerize and run

Create Dockerfile next to app.py:


FROM python:3.12-slim
WORKDIR /app
COPY app.py .
USER 10001
EXPOSE 8080
CMD ["python", "app.py"]

Build and run the image:


docker build -t learning-service:local .
docker run --rm --name learning-service -p 8080:8080 learning-service:local

In another terminal, run curl -fsS http://localhost:8080/. The expected response is hello from the service. Stop the container with Ctrl+C before continuing. For more context, see the Docker notes.

Deploy to local Kubernetes

Create a cluster and load the image into it:


kind create cluster --name learning-service
kind load docker-image learning-service:local --name learning-service

Create k8s.yaml:


apiVersion: apps/v1
kind: Deployment
metadata:
  name: learning-service
spec:
  replicas: 1
  selector:
    matchLabels:
      app: learning-service
  template:
    metadata:
      labels:
        app: learning-service
    spec:
      containers:
        - name: service
          image: learning-service:local
          imagePullPolicy: Never
          ports:
            - containerPort: 8080
          readinessProbe:
            httpGet:
              path: /healthz
              port: 8080
          livenessProbe:
            httpGet:
              path: /healthz
              port: 8080
          resources:
            requests:
              cpu: 25m
              memory: 32Mi
            limits:
              cpu: 250m
              memory: 128Mi

Apply it and wait for readiness:


kubectl apply -f k8s.yaml
kubectl rollout status deployment/learning-service
kubectl port-forward deployment/learning-service 8080:8080

In another terminal, curl -fsS http://localhost:8080/healthz should again return ok. The kubectl rollout status command should report a successful rollout. If it does not, inspect kubectl get pods, kubectl describe pod <pod-name>, and kubectl logs deployment/learning-service. The Kubernetes concepts page explains the objects involved.

Change and verify

Stop port forwarding. Change the / response in app.py, rebuild the image, then load it into kind again. Restart the Deployment and check the new response:


docker build -t learning-service:local .
kind load docker-image learning-service:local --name learning-service
kubectl rollout restart deployment/learning-service
kubectl rollout status deployment/learning-service
kubectl port-forward deployment/learning-service 8080:8080

In another terminal, fetch / and confirm it returns your new response. Think about the gap between “the Deployment is ready” and “users can complete their task.” Which additional check would close that gap? Continue with the SRE path or CI feedback loop field note.

Clean up

Stop port forwarding, then run kind delete cluster --name learning-service. This removes the local cluster created for the lab.