This directory contains the Kubernetes manifests and GitHub workflow for deploying the Babbly application infrastructure to Kubernetes clusters.

The infrastructure includes the following components:

1. Kubernetes Cluster:

   • Any Kubernetes cluster (AKS, EKS, GKE, or local)
   • Sufficient resources for all components

2. GitHub Secrets (for automated deployment): Set up the following secret in your GitHub repository:

   KUBE_CONFIG_DATA - Base64 encoded kubeconfig file content
   

   To get your kubeconfig:

   # Get cluster credentials (example for AKS)
   az aks get-credentials --resource-group your-resource-group --name your-aks-cluster
   
   # Base64 encode the config
   cat ~/.kube/config | base64 | tr -d '\n'

   Copy the output and paste it as the KUBE_CONFIG_DATA secret value in GitHub.

Infrastructure Workflow (this repo): The GitHub workflow deploys ONLY infrastructure components:

• Secrets (Auth0, Cassandra credentials)
• Zookeeper
• Kafka
• Cassandra
• Prometheus (monitoring)
• Grafana (dashboards)

Microservices (separate workflows): Each microservice has its own CI/CD pipeline in their respective repositories:

• Auth Service - deployed via babbly-auth-service/release.yml
• User Service - deployed via babbly-user-service/release.yml
• Post Service - deployed via babbly-post-service/release.yml
• Comment Service - deployed via babbly-comment-service/release.yml
• Like Service - deployed via babbly-like-service/release.yml
• API Gateway - deployed via babbly-api-gateway/release.yml
• Frontend - deployed via babbly-frontend/release.yml

If you need to deploy infrastructure manually (run from repository root):

# Deploy infrastructure components
kubectl apply -f infrastructure/k8s/secrets.yaml --namespace=default
kubectl apply -f infrastructure/k8s/kafka-zookeeper.yaml --namespace=default
kubectl apply -f infrastructure/k8s/cassandra.yaml --namespace=default
kubectl apply -f infrastructure/k8s/prometheus.yaml --namespace=default
kubectl apply -f infrastructure/k8s/grafana.yaml --namespace=default

# Wait for infrastructure to be ready
kubectl wait --for=condition=available --timeout=300s deployment/kafka --namespace=default

# Cassandra takes longer to start, so wait for StatefulSet first
kubectl wait --for=jsonpath='{.status.readyReplicas}'=1 --timeout=300s statefulset/cassandra --namespace=default
kubectl wait --for=condition=ready --timeout=600s pod -l app=cassandra --namespace=default

kubectl wait --for=condition=available --timeout=300s deployment/prometheus --namespace=default
kubectl wait --for=condition=available --timeout=300s deployment/grafana --namespace=default

Note: Microservices are deployed separately via their individual CI/CD workflows when code changes are pushed to each service repository.

After deployment:

1. Get service information:

   kubectl get services

2. Frontend: Access via NodePort or configured ingress

3. API Gateway: Access via LoadBalancer external IP or ingress

4. Prometheus: Access via NodePort or port-forward

5. Grafana: Access via NodePort or port-forward

   • Username: admin
   • Password: ${GRAFANA_ADMIN_PASSWORD}

• PostgreSQL: Using external cloud service (configurable)
• Cassandra: Deployed in cluster with persistent storage
• Keyspaces: babbly_posts, babbly_comments, babbly_likes

• Kafka Topics:
  • user-events
  • auth-events
  • post-events
  • comment-events
  • like-events

Services communicate using Kubernetes DNS:

• auth-service
• user-service
• post-service
• comment-service
• like-service
• kafka
• cassandra

• Prometheus: Collects metrics from all Babbly services and Kubernetes cluster

  • Configured to scrape /metrics endpoints from all microservices
  • Monitors Kubernetes nodes, pods, and services
  • Persistent storage with configurable retention

• Grafana: Provides visualization dashboards

  • Pre-configured Prometheus datasource
  • Built-in dashboards:
    • Babbly Application Overview: Service health, HTTP request rates, response times
    • Kubernetes Cluster Overview: Pod status, CPU/memory usage
  • Supports custom dashboard creation

Cassandra Startup Issues: If Cassandra readiness probe fails after configuration changes:

# Check Cassandra status
kubectl get statefulset cassandra --namespace=default
kubectl get pods -l app=cassandra --namespace=default

# If probe config didn't update, restart the pod
kubectl delete pod cassandra-0 --namespace=default

# Wait for it to come back up
kubectl wait --for=condition=ready --timeout=600s pod -l app=cassandra --namespace=default

# Check logs if still having issues
kubectl logs cassandra-0 --namespace=default --tail=100

kubectl get pods --namespace=default
kubectl describe pod <pod-name> --namespace=default
kubectl logs <pod-name> --namespace=default

kubectl get services --namespace=default
kubectl get endpoints --namespace=default

kubectl exec deployment/kafka --namespace=default -- kafka-topics --bootstrap-server localhost:9092 --list

kubectl exec cassandra-0 --namespace=default -- cqlsh -e "DESCRIBE KEYSPACES;"

# Check Prometheus targets
kubectl port-forward service/prometheus ${PROMETHEUS_PORT}:${PROMETHEUS_PORT} --namespace=default
# Then visit http://localhost:${PROMETHEUS_PORT}/targets

# Check Grafana dashboards
kubectl port-forward service/grafana ${GRAFANA_PORT}:${GRAFANA_PORT} --namespace=default
# Then visit http://localhost:${GRAFANA_PORT} (admin/${GRAFANA_ADMIN_PASSWORD})

# Check Prometheus metrics collection
kubectl logs deployment/prometheus --namespace=default

# Check Grafana logs
kubectl logs deployment/grafana --namespace=default

All services are configured for development environment:

• ASPNETCORE_ENVIRONMENT=Development
• Detailed logging enabled
• CORS configured for local development
• Default database credentials (change for production)

For Prometheus monitoring to work properly, each microservice should expose metrics at /metrics endpoint. Consider adding:

• ASP.NET Core health checks (/health)
• Prometheus metrics middleware
• Custom business metrics (posts created, users registered, etc.)
• HTTP request/response metrics
• Database connection metrics

⚠️ For production deployment:

• Change all default credentials before deployment
• Use strong passwords and rotate them regularly
• Use proper TLS certificates
• Configure network policies
• Use secure secret management (Azure Key Vault, AWS Secrets Manager, etc.)
• Enable RBAC and pod security policies
• Restrict Prometheus scraping permissions
• Configure Grafana OAuth integration
• Implement proper authentication and authorization
• Regular security audits and updates