Event Driven Microservices

# Event-Driven Microservices Architecture

## Objective

Design and implement a production-grade event-driven microservices architecture using message brokers (Kafka/RabbitMQ) with Saga pattern for distributed transactions. The system must be resilient, scalable, and maintain data consistency across service boundaries.

---

## Core Architecture Principles

### Mandatory Requirements

1. **Event-Driven Communication**: Services communicate exclusively through asynchronous events, never direct HTTP calls between services
2. **Saga Pattern**: Implement choreography-based or orchestration-based Saga for distributed transactions
3. **Event Sourcing**: Store state changes as immutable event log for audit and replay capabilities
4. **CQRS**: Separate read and write models for optimal performance and scalability
5. **Idempotency**: All event handlers must be idempotent to handle duplicate messages

---

## Technology Stack

### Message Broker Selection

**Apache Kafka** (High-throughput, event streaming):
- Use for: Event sourcing, real-time analytics, high-volume event streams
- Retention: Configure based on replay requirements (7-30 days typical)
- Partitioning: Partition by aggregate ID for ordering guarantees

**RabbitMQ** (Flexible routing, traditional messaging):
- Use for: Task queues, RPC patterns, complex routing scenarios
- Exchanges: Topic exchange for event broadcasting, direct exchange for commands
- Dead Letter Queues: Configure DLQ for failed message handling

### Core Components

- **Event Store**: EventStoreDB or custom implementation on PostgreSQL
- **Schema Registry**: Confluent Schema Registry or AWS Glue for event schema versioning
- **Service Mesh**: Istio or Linkerd for observability and traffic management
- **Distributed Tracing**: Jaeger or Zipkin for request flow visualization
- **Monitoring**: Prometheus + Grafana for metrics, ELK stack for logs

---

## Event Design Standards

### Event Structure