Skip to content

README.md

Latest commit

 

History

History
67 lines (47 loc) · 4.08 KB

README.md

File metadata and controls

67 lines (47 loc) · 4.08 KB
Raw

Stream-Seat: Cloud-Native Microservices Migration

Project Overview

Stream-Seat is a distributed microservices platform designed for real-time ticket booking. This project documents the transition from a localized development environment to a production-grade, serverless AWS Cloud infrastructure, focusing on high availability and secure service orchestration.

Cloud Architecture (The "100% Infrastructure Milestone")

The project has successfully reached full infrastructure deployment, transitioning from manual task management to a fully routed, discovery-aware microservices ecosystem.

Current Infrastructure Components:

  • Compute: AWS ECS Fargate (Serverless) utilizing the AMD64 runtime environment.
  • Service Discovery: AWS Cloud Map with Private DNS Namespaces (stream-seat.local) for stable, hostname-based internal routing.
  • API Gateway: Spring Cloud Gateway deployed in public subnets as the unified entry point for all external traffic.
  • Database: Amazon RDS (MySQL 8.0) isolated in private subnets with administrative access via EC2 Bastion Hosts.
  • Networking: Custom VPC with multi-AZ redundancy and Security Group Nesting for zero-trust inter-service communication.

🛠️ Technical Challenges & Solutions

1. Cross-Platform Container Compatibility

  • Issue: CannotPullContainerError due to ARM64 (Apple Silicon) vs. AMD64 (AWS Fargate) architecture mismatch.
  • Fix: Implemented Docker multi-platform builds using the --platform linux/amd64 flag to ensure binary compatibility across development and cloud environments.

2. Private Database Orchestration & Schema Initialization

  • Issue: Security isolation prevented RDS schema creation, causing SQLGrammarException during service startup.
  • Fix: Utilized a Bastion Host (EC2) to bridge into the private subnet. Manually initialized the stream_seat_db schema to support Hibernate's DDL execution while maintaining a restricted database attack surface.

3. Non-Routable Link-Local Address Conflict

  • Issue: Fargate tasks registered with Eureka using 169.254.x.x addresses, which are non-routable between different tasks, causing "Connection Refused" errors at the Gateway.
  • Fix: Configured Spring Cloud InetUtils via environment variables (SPRING_CLOUD_INETUTILS_PREFERRED_NETWORKS=10.0) to force services to register using their routable VPC Private IPs.

4. Ephemeral IP Management

  • Issue: Dynamic Fargate IPs broke hardcoded service links on every task restart, leading to frequent manual configuration updates.
  • Fix: Integrated AWS Cloud Map to provide a persistent Private DNS record (eureka.stream-seat.local), decoupling the registry location from its ephemeral IP address.

🚦 Deployment Status

Service Status Port Networking Registry
Discovery Server (Eureka) RUNNING 8761 Public Subnet ECR (us-east-1)
API Gateway RUNNING 8082 Public Subnet ECR (us-east-1)
Booking Service RUNNING 8080 Private Subnet ECR (us-east-1)
Event Service RUNNING 8081 Private Subnet ECR (us-east-1)
RDS MySQL AVAILABLE 3306 Private Subnet AWS RDS

📈 Future Roadmap (CI/CD)

  • GitHub Actions Integration: Automate the build-test-push cycle for all microservices.
  • Automated Deployment: Configure ECS task definitions to update automatically upon new image pushes.
  • Secret Management: Move AWS credentials and database passwords to GitHub Secrets and AWS Secrets Manager.

🛠️ Key Skills Demonstrated

  • Cloud Orchestration: Managing complex service lifecycles via AWS ECS and Cloud Map.
  • Network Security: Designing multi-tier VPCs and implementing Bastion Host architectures.
  • Infrastructure Troubleshooting: Debugging Link-Local networking and cross-architecture container issues.
  • Externalized Configuration: Utilizing Spring Boot environment variables to override production settings dynamically without image rebuilds.