AWS EventBridge

What it is

Amazon EventBridge is a fully managed serverless event bus that facilitates building scalable, decoupled event-driven architectures. It acts as an intelligent intermediary that receives events from multiple sources — AWS services, custom applications, and dozens of native SaaS integrations — and routes them to specific destinations based on sophisticated content-based rules.

Unlike traditional messaging systems that require queue or topic configuration, EventBridge uses declarative event patterns that enable granular filtering without additional code. Events are processed asynchronously with delivery guarantees and automatic retry capabilities, eliminating the need to manage messaging infrastructure.

EventBridge integrates natively with the AWS ecosystem and provides a Schema Registry that auto-discovers and versions event structures, facilitating contract evolution between services in microservices architectures.

Core components

Production event patterns

{
  "version": "0",
  "id": "7bf73129-1428-4cd3-a780-95db273d1602",
  "detail-type": "Order Status Changed",
  "source": "ecommerce.orders",
  "account": "123456789012",
  "time": "2024-03-19T10:30:00Z",
  "region": "us-east-1",
  "detail": {
    "orderId": "order-12345",
    "previousStatus": "processing",
    "currentStatus": "shipped",
    "customerId": "cust-67890",
    "amount": 149.99,
    "items": [
      {
        "sku": "PROD-001",
        "quantity": 2,
        "price": 74.99
      }
    ]
  }
}

Advanced filtering rule

{
  "source": ["ecommerce.orders"],
  "detail-type": ["Order Status Changed"],
  "detail": {
    "currentStatus": ["shipped", "delivered"],
    "amount": [{"numeric": [">", 100]}],
    "customerId": [{"exists": true}]
  }
}

EventBridge Pipes vs Rules

EventBridge Pipes introduce a point-to-point integration model that complements the traditional pub/sub model of EventBridge Rules:

Event archive and replay

EventBridge provides native archive and replay capabilities that are critical for production systems:

• Archive: Automatically stores events for a configurable retention period (indefinite by default)
• Replay: Re-processes historical events to specific targets
• Use cases: Debugging, testing with real data, failure recovery, retrospective analysis

# Create archive for order events
aws events put-archive \
  --archive-name orders-archive \
  --event-source-arn arn:aws:events:us-east-1:123456789012:event-bus/orders-bus \
  --retention-days 365
 
# Replay events from last month
aws events start-replay \
  --replay-name orders-replay-march \
  --event-source-arn arn:aws:events:us-east-1:123456789012:archive/orders-archive \
  --event-start-time 2024-03-01T00:00:00Z \
  --event-end-time 2024-03-31T23:59:59Z

Decision: EventBridge vs SQS vs SNS

Decision framework

Use EventBridge when:

• You need content-based event routing
• Integrating multiple AWS services and SaaS
• You require schema registry and versioning
• Building complex event-driven architectures
• Second-level latency is acceptable

Use SQS when:

• You need strict FIFO guarantees
• You require sub-second latency
• Implementing work queues or task processing
• Cost per message is critical

Use SNS when:

• Implementing simple notifications
• You need basic fan-out without filtering
• Integrating with email, SMS, or mobile push
• You require minimal latency

Observability integration

EventBridge integrates natively with CloudWatch and AWS X-Ray to provide complete visibility:

• Rule metrics: Invocations, MatchedEvents, FailedInvocations — monitor event processing volume and detect target failures
• Delivery metrics: InvocationsSentToDeadLetterQueue, ThrottledRules — identify undelivered events and throttling
• X-Ray tracing: End-to-end tracing of event flow from source to final target

Common anti-patterns

• Single event bus: Using the default bus for all domains reduces isolation and complicates governance
• Synchronous events: EventBridge is asynchronous; don't use for cases requiring immediate response
• Large payloads: 256KB limit; use S3 for large data and reference in event
• Over-filtering: Overly specific rules create coupling; prefer filtering in targets when possible

Why it matters

EventBridge represents a fundamental shift toward serverless event-driven architectures that scale automatically and reduce operational coupling. For engineering teams, it eliminates the complexity of managing message brokers while providing advanced capabilities like schema registry, declarative transformations, and native SaaS integrations. The archive and replay capability is especially valuable for critical systems where event auditability and recovery are business requirements. The per-processed-event pricing model aligns costs with generated value, making incremental adoption viable in existing systems.

References

• Amazon EventBridge User Guide — AWS, 2024. Complete official documentation.
• EventBridge Event Patterns — AWS, 2024. Filtering pattern syntax and examples.
• EventBridge Pipes Documentation — AWS, 2024. Point-to-point integration guide.
• Serverless Land EventBridge Patterns — AWS, 2024. Collection of architectural patterns.
• EventBridge Schema Registry — AWS, 2024. Schema management and versioning.
• Event-Driven Architecture — AWS, 2024. Architectural guides and best practices.