This repository demonstrates a real-world migration of a monolithic application to a microservices architecture using the Strangler Fig Pattern and Change Data Capture (CDC).

The project simulates a transformation from a single Go monolith with a shared database to independent microservices, each with its own isolated database, kept in sync via Kafka and Debezium.

• Source: Go Monolith (Users & Orders) + PostgreSQL
• Streaming Backbone: Kafka + ZooKeeper + Debezium
• Target Microservice 1: User Service (Go) + Isolated PostgreSQL
• Target Microservice 2: Order Service (Go) + Isolated PostgreSQL
• Entry Point: API Gateway (Go) implementing Strangler Fig Fallback

• /monolith: The legacy application (Source of Truth).
• /api-gateway: The "Strangler" entry point with transparent fallback logic.
• /user-service: The new decomposed User microservice.
• /order-service: The new decomposed Order microservice.
• /traffic-generator: Simulates real-world load on the monolith.
• /stream-consumer: A verification tool for Kafka events.
• /docs: Detailed guides for each migration phase.

1. Phase 1: Foundation - Baseline monolith and infrastructure.
2. Phase 2: CDC Pipeline - Turning the DB into an event source.
3. Phase 3: Decomposition - Implementing the User service and data sync.
4. Phase 4: Scaling Out - Implementing the Order service.
5. Phase 5: Traffic Cutover - Implementing the API Gateway and Strangler Pattern.
6. Phase 6: Historical Data Backfill - Bulk migration utility.
7. Phase 7: Migration Observability - Real-time Mission Control Dashboard.

• Language: Go 1.21+
• Databases: PostgreSQL 15, SQLite (optional for quick tests)
• Message Broker: Kafka (Confluent Platform)
• CDC Tool: Debezium Connect
• Deployment: Docker Compose

1. Start the infrastructure:

   docker compose up -d