deploy-kubernetes.md

Kubernetes Deployment

The distributed mode can also be deployed on Kubernetes, using Spring Cloud Kubernetes for service discovery instead of Eureka.

Table of Contents

• Deploying to Local Kubernetes (Kind)
  • Prerequisites
  • Build and Load Images
  • Deploy to Kubernetes
  • Access the Application
  • Access Jaeger UI
  • Cleanup
• Azure Kubernetes Service (AKS) Deployment
  • Prerequisites
  • Setup AKS Cluster
  • Register Azure Resource Providers (One-time setup)
  • Push Images to Azure Container Registry
  • Deploy to Azure
  • Cleanup Azure Resources
  • Managing Multiple Clusters (Local vs Azure)

Deploying to Local Kubernetes (Kind)

Prerequisites

Install the following packages:

• Docker - Container runtime
• kubectl - Kubernetes CLI
• Kind (recommended) – Local Kubernetes cluster

Create a Kind cluster:

kind create cluster --name microservices

Build and Load Images

# Build all Docker images
docker compose build

# Load images into Kind cluster
for img in simulator quizzes; do
  kind load docker-image ${img}:latest --name microservices
done

Deploy to Kubernetes

# Create namespace and RBAC
kubectl apply -f k8s/namespace.yaml
kubectl apply -f k8s/rbac.yaml
kubectl apply -f k8s/configmap.yaml

# Deploy infrastructure
kubectl apply -f k8s/infrastructure/rabbitmq.yaml
kubectl apply -f k8s/infrastructure/jaeger.yaml

# Wait for infrastructure to be ready
kubectl wait --for=condition=ready pod -l app=rabbitmq -n microservices-simulator --timeout=120s
kubectl wait --for=condition=ready pod -l app=jaeger -n microservices-simulator --timeout=60s

# Deploy microservices (choose one)
# For stream communication
kubectl apply -f k8s/services-stream/

# For gRPC communication
# kubectl apply -f k8s/services-grpc/

# Check status
kubectl get pods -n microservices-simulator

Note: To change transactional model profile, edit k8s/services-stream/ or k8s/services-grpc/ and change the SPRING_PROFILES_ACTIVE environment variable of each service.

Access the Application

# Port-forward to gateway
kubectl port-forward svc/gateway 8080:8080 -n microservices-simulator

(For all default endpoints including rabbitmq and service discovery, see the Service URLs and Ports section in the main README).

Access Jaeger UI

kubectl port-forward svc/jaeger 16686:16686 -n microservices-simulator

Then open http://localhost:16686 to view distributed traces.

Cleanup

kubectl delete namespace microservices-simulator

---

Azure Kubernetes Service (AKS) Deployment

Deploy the distributed mode to Azure Kubernetes Service for cloud-based deployments.

Prerequisites

• Azure CLI installed
• Active Azure subscription (e.g., Azure for Students)

Setup AKS Cluster

# Login to Azure
az login

# Create Resource Group
az group create --name simulator-rg-es --location spaincentral

# Create AKS Cluster (Free tier, minimal resources) -- This is a cluster example
az aks create \
  --resource-group simulator-rg-es \
  --name simulator-cluster \
  --tier free \
  --node-count 1 \
  --node-vm-size Standard_B2s_v2 \
  --generate-ssh-keys

# Connect to the Cluster
az aks get-credentials --resource-group simulator-rg-es --name simulator-cluster

# Verify connection
kubectl get nodes

Register Azure Resource Providers (One-time setup)

# Register Container Registry provider (required for ACR)
az provider register --namespace Microsoft.ContainerRegistry

# Check registration status (wait until "Registered")
az provider show --namespace Microsoft.ContainerRegistry --query "registrationState"

Push Images to Azure Container Registry

# Run the push script (creates ACR, attaches to AKS, pushes images)
chmod +x scripts/push-to-acr.sh
./scripts/push-to-acr.sh

Deploy to Azure

# 1. Base setup
kubectl apply -f k8s/namespace.yaml
kubectl apply -f k8s/configmap.yaml
kubectl apply -f k8s/rbac.yaml

# 2. Infrastructure (Centralized PostgreSQL + RabbitMQ)
kubectl apply -f k8s/infrastructure/

# 3. Wait for infrastructure to be ready
kubectl wait --for=condition=ready pod -l app=postgres -n microservices-simulator --timeout=180s
kubectl wait --for=condition=ready pod -l app=rabbitmq -n microservices-simulator --timeout=120s

# 4. Deploy Azure-optimized microservices (uses ACR images + centralized DB)
kubectl apply -f k8s/services-azure/

# 5. Check status
kubectl get pods -n microservices-simulator

# 6. Access the gateway
kubectl get svc gateway -n microservices-simulator
# Or use port-forward
kubectl port-forward -n microservices-simulator svc/gateway 8080:8080

Save costs by stopping the cluster when not in use:

# Stop the cluster
az aks stop --name simulator-cluster --resource-group simulator-rg-es

# Start the cluster again
az aks start --name simulator-cluster --resource-group simulator-rg-es

Cleanup Azure Resources

# Delete the cluster
az aks delete --name simulator-cluster --resource-group simulator-rg-es

# Delete everything (including ACR)
az group delete --name simulator-rg-es

Managing Multiple Clusters (Local vs Azure)

When using both local (Kind) and Cloud (Azure) clusters, your kubectl context may be pointing to the wrong cluster.

To see all available clusters:

kubectl config get-contexts

To switch back to your local Kind cluster:

kubectl config use-context kind-microservices

To switch to your Azure AKS cluster:

kubectl config use-context simulator-cluster