Passion Project: Multi-Service Helm Chart

After this lesson, you will be able to: Deploy a multi-service application to a local Kubernetes cluster using Helm, with resource limits, health checks, and a rolling update strategy.

Passion project for the K8s sub-track per Curriculum-Upgrade.md. Take the multi-service stack from do-docker-passion (or any equivalent) and ship it to K8s with a Helm chart you wrote.

Prerequisites:Observability

Milestone 1 — Spin up a real cluster

kind is fine; GKE Autopilot free-trial is even better.

1
Local: kind create cluster --config kind-config.yaml (config below)
2
Cloud (optional): GKE Autopilot free trial gives a real LB + cert provisioning
3
Verify: kubectl get nodes (Ready)
4
Install add-ons via Helm: ingress-nginx, cert-manager (if cloud), kube-prometheus-stack

kind-config.yaml (extras for Ingress)

Map host ports so Ingress is reachable from your laptop browser.

tsx

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
  - role: control-plane
    kubeadmConfigPatches:
      - |
        kind: InitConfiguration
        nodeRegistration:
          kubeletExtraArgs:
            node-labels: "ingress-ready=true"
    extraPortMappings:
      - containerPort: 80
        hostPort: 80
      - containerPort: 443
        hostPort: 443

Milestone 2 — Author the Helm chart

helm create scaffolds the template.

1
helm create myapp
2
Customize templates/deployment.yaml: add real probes (readinessProbe + livenessProbe)
3
Add templates/service.yaml: ClusterIP for app + StatefulSet/Service for Postgres
4
Add templates/ingress.yaml: route /api → api service, / → web service
5
Add templates/configmap.yaml + templates/secret.yaml for env vars
6
values.yaml has every config: image tags, replica counts, resources, hostname
7
Validate: helm template ./myapp ; helm lint ./myapp

Milestone 3 — Deploy

One-command install. Then verify everything.

1
helm install myapp ./myapp -n myapp --create-namespace
2
Watch: kubectl get pods -n myapp -w
3
Wait for all pods Ready (probes succeed)
4
Verify Service: kubectl get svc -n myapp
5
Verify Ingress: kubectl get ingress -n myapp
6
Browse to the host (localhost in kind; the LB IP in cloud)
7
Watch a rollout: kubectl set image deployment/web web=myapp:1.1 -n myapp ; kubectl rollout status
8
Rollback: helm rollback myapp -n myapp

Milestone 4 — Documentation

README + architecture diagram. This is the artefact.

1
README.md at the chart root: how to install, what values to set, what the architecture is
2
Architecture diagram: services, Ingress, Service mesh (if any), DB. Excalidraw → PNG.
3
For each Pod: what's its job, what does it depend on, what are its probes
4
Link the live URL (if using cloud)
5
License + Contributing

Common mistakes only candidates with offers avoid

Hard-coded values in templates. Defeats Helm. Missing readiness probes. The rolling deploy drops requests. Database in K8s without a backup story. Don't claim 'production-grade' without one. Skipping the architecture diagram. Hiring managers skim READMEs. No verification step, never running `helm template` before applying. Big YAML, easy typo.

←Observability

Back to Kubernetes

Kubernetes Job Readiness→