Project 3: Microservices on AKS with Azure Container Registry, Bicep and GitHub Actions CI/CD
What We Build
text
GitHub repo -> Actions: build image -> push to ACR -> deploy to AKS
|
AKS (3 nodes, zones) : web (frontend) -> orders-api -> Azure SQL / Cosmos DB (private)
Ingress controller + public IP, Container insights to Log Analytics, ACR attached to AKS (AcrPull)
Step 1: Infrastructure with Bicep
bicep
param location string = resourceGroup().location
param suffix string = uniqueString(resourceGroup().id)
resource acr 'Microsoft.ContainerRegistry/registries@2023-07-01' = {
name: 'acr${suffix}'
location: location
sku: { name: 'Standard' }
properties: { adminUserEnabled: false }
}
resource law 'Microsoft.OperationalInsights/workspaces@2022-10-01' = {
name: 'law-aks-${suffix}'
location: location
properties: { sku: { name: 'PerGB2018' }, retentionInDays: 30 }
}
resource aks 'Microsoft.ContainerService/managedClusters@2024-05-01' = {
name: 'aks-shop'
location: location
identity: { type: 'SystemAssigned' }
properties: {
dnsPrefix: 'aksshop'
agentPoolProfiles: [{
name: 'system', count: 3, vmSize: 'Standard_D2s_v5', mode: 'System',
availabilityZones: ['1', '2', '3'], enableAutoScaling: true, minCount: 3, maxCount: 6
}]
oidcIssuerProfile: { enabled: true }
securityProfile: { workloadIdentity: { enabled: true } }
addonProfiles: { omsagent: { enabled: true, config: { logAnalyticsWorkspaceResourceID: law.id } } }
}
}
output acrName string = acr.name
output acrLogin string = acr.properties.loginServer
bash
az group create -n rg-aks -l centralindia
az deployment group create -g rg-aks -f aks.bicep
az aks update -g rg-aks -n aks-shop --attach-acr <acrName> # lets AKS pull images (AcrPull)
az aks get-credentials -g rg-aks -n aks-shop
kubectl get nodes
Step 2: App and Dockerfile
dockerfile
FROM python:3.12-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
USER 10001
EXPOSE 8080
CMD ["gunicorn", "-b", "0.0.0.0:8080", "app:app"]
bash
az acr build -r <acrName> -t orders-api:v1 ./orders-api # build in the cloud, no local Docker needed
az acr repository list -n <acrName> -o table
Step 3: Kubernetes Manifests
yaml
apiVersion: apps/v1
kind: Deployment
metadata: { name: orders-api }
spec:
replicas: 3
selector: { matchLabels: { app: orders-api } }
template:
metadata: { labels: { app: orders-api } }
spec:
containers:
- name: orders-api
image: <acrName>.azurecr.io/orders-api:v1
ports: [{ containerPort: 8080 }]
resources: { requests: { cpu: 100m, memory: 128Mi }, limits: { cpu: 500m, memory: 256Mi } }
readinessProbe: { httpGet: { path: /health, port: 8080 }, initialDelaySeconds: 5 }
livenessProbe: { httpGet: { path: /health, port: 8080 }, initialDelaySeconds: 15 }
---
apiVersion: v1
kind: Service
metadata: { name: orders-api }
spec:
selector: { app: orders-api }
ports: [{ port: 80, targetPort: 8080 }]
---
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata: { name: orders-api }
spec:
scaleTargetRef: { apiVersion: apps/v1, kind: Deployment, name: orders-api }
minReplicas: 3
maxReplicas: 10
metrics: [{ type: Resource, resource: { name: cpu, target: { type: Utilization, averageUtilization: 70 } } }]
bash
kubectl apply -f k8s/
kubectl get pods,svc,hpa
az aks approuting enable -g rg-aks -n aks-shop # managed NGINX ingress (check current add-on guidance)
kubectl apply -f k8s/ingress.yaml
Step 4: Workload Identity for Secrets and Data
- Create a user-assigned managed identity, add a federated credential linked to the Kubernetes service account
- Give the identity a role such as Key Vault Secrets User or database access
- Annotate the service account and pod. The app then uses
DefaultAzureCredentialwith no secrets
Step 5: GitHub Actions Pipeline
yaml
ame: build-and-deploy
on: { push: { branches: [main] } }
permissions: { id-token: write, contents: read }
env:
ACR: myacr
IMAGE: orders-api
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: azure/login@v2
with:
client-id: ${{ secrets.AZURE_CLIENT_ID }}
tenant-id: ${{ secrets.AZURE_TENANT_ID }}
subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
- name: Build and push image
run: az acr build -r $ACR -t $IMAGE:${{ github.sha }} ./orders-api
- uses: azure/aks-set-context@v4
with: { resource-group: rg-aks, cluster-name: aks-shop }
- name: Deploy
run: |
kubectl set image deployment/orders-api orders-api=$ACR.azurecr.io/$IMAGE:${{ github.sha }}
kubectl rollout status deployment/orders-api --timeout=120s
- Tag images with the commit SHA, not
latest, so releases are traceable and can be rolled back withkubectl rollout undo - Add steps for tests, image scanning (Defender for Containers) and deploy to staging before production
Step 6: Observe and Test
- Open Container insights to see node and pod CPU and memory
- Kill a pod with
kubectl delete podand watch the Deployment recreate it - Load test and watch the HPA add pods, and the cluster autoscaler add nodes
- Push a code change and watch the pipeline roll it out with zero downtime
Review
| Pillar | Notes |
|---|---|
| Reliability | 3 zones, probes, replicas, rolling updates, backups for stateful data |
| Security | Workload identity, ACR with no admin user, private cluster option, Azure Policy for Kubernetes, image scanning |
| Cost | Autoscaling, Spot node pools for batch, right-size requests and limits, stop dev clusters |
| Operations | Bicep, GitHub Actions, Container insights, upgrade channels |
| Performance | HPA, cluster autoscaler, caching, node sizing |
Clean Up
bash
az group delete -n rg-aks --yes --no-wait
Conclusion
You deployed containers to AKS with infrastructure in Bicep, images in ACR, manifests with probes and autoscaling, workload identity and a GitHub Actions pipeline using federated login.
Quick Summary: Bicep; image; manifests; identity; pipeline; observe; review; cleanup.

