event-driven

Event-Driven Architecture (EDA)

EDA is a software architecture paradigm promoting the production, detection, consumption of, and reaction to events. In 2025, it is the backbone of real-time, scalable, and decoupled systems.

When to Use

• When strict decoupling is required (Producer doesn't know Consumer).
• High-volume, bursty traffic (using buffering queues).
• Asynchronous workflows (e.g., "User Signed Up" -> Send Email, Create Wallet, Analytics).
• Real-time updates (WebSockets/Push).

Quick Start

// Producer (Order Service)
await messageBroker.publish("order.created", {
  orderId: "123",
  userId: "456",
  timestamp: Date.now(),
});

// Consumer (Shipping Service)
// Doesn't need to be online when order is created
messageBroker.subscribe("order.created", async (event) => {
  await shippingService.schedulePickup(event.orderId);
  console.log("Shipping scheduled");
});

// Consumer (Analytics Service)
// New feature added later? No changes to Order Service!
messageBroker.subscribe("order.created", async (event) => {
  await analytics.trackRevenue(event);
});

Core Concepts

Event

A significant change in state (Immutable fact). "OrderCreated" not "CreateOrder".

Broker (Event Bus)

The middleware (Kafka, RabbitMQ, SNS/SQS) that receives, stores, and routes events.

Pub/Sub

Pattern where publishers send messages to a topic, and multiple subscribers receive them independently.

Common Patterns

Event Sourcing

Storing the state of an entity as a sequence of state-changing events rather than the current snapshot.

CQRS (Command Query Responsibility Segregation)

Separating the Read and Write models. Writes publish events; Reads update a denormalized view based on those events.

Transactional Outbox

Ensuring data consistency. Write the event to a DB table in the same transaction as the business logic, then a background worker pushes it to the broker.

Best Practices

Do:

• Use Schemas (Avro, Protobuf, JSON Schema) to govern event structure (Schema Registry).
• Ensure Idempotency in consumers (handling the same message twice safely).
• Monitor Lag (how far behind consumers are).