Skip to main content

Event-Driven Processing

Azotte uses event-driven architecture to enable real-time processing, loose coupling, and scalable system integration.

Architecture Overview​

┌───────────────┐     ┌───────────────────────┐     ┌─────────────────┐
│ Event Producers │────▶│     Message Bus      │────▶│ Event Consumers │
│                 │     │                       │     │                 │
│• Subscription   │     │• Apache Kafka       │     │• Webhook Service │
│• Payment        │     │• RabbitMQ          │     │• Analytics       │
│• Campaign       │     │• Event Store       │     │• Notifications   │
│• Customer       │     │• Message Routing   │     │• Integrations    │
└───────────────┘     └───────────────────────┘     └─────────────────┘

Event Types​

Subscription Events​

  • subscription.created - New subscription activated
  • subscription.updated - Subscription details changed
  • subscription.cancelled - Subscription cancelled
  • subscription.renewed - Subscription renewed
  • subscription.paused - Subscription temporarily paused
  • subscription.resumed - Subscription reactivated

Payment Events​

  • payment.authorized - Payment successfully authorized
  • payment.captured - Payment captured/settled
  • payment.failed - Payment attempt failed
  • payment.refunded - Refund processed
  • payment.disputed - Chargeback/dispute initiated

Customer Events​

  • customer.created - New customer registered
  • customer.updated - Customer profile changed
  • customer.deleted - Customer account deleted

Campaign Events​

  • campaign.activated - Campaign went live
  • campaign.applied - Campaign applied to purchase
  • campaign.expired - Campaign reached end date

Event Schema​

Standard Event Structure​

{
  "id": "evt_1234567890",
  "type": "subscription.created",
  "version": "1.0",
  "timestamp": "2024-03-14T10:30:00Z",
  "tenant_id": "tenant_123",
  "correlation_id": "corr_abc123",
  "data": {
    "subscription": {
      "id": "sub_1234567890",
      "customer_id": "cus_1234567890",
      "status": "active",
      "bundle_id": "bundle_premium",
      "price_id": "price_monthly_99"
    }
  },
  "metadata": {
    "source": "subscription-service",
    "idempotency_key": "idem_xyz789"
  }
}

Event Versioning​

  • Semantic versioning for event schemas
  • Backward compatibility guarantees
  • Gradual migration strategies
  • Schema evolution support

Message Bus Infrastructure​

Apache Kafka​

  • Use Case: High-throughput event streaming
  • Benefits: Durability, scalability, ordering
  • Topics: Partitioned by tenant or event type

RabbitMQ​

  • Use Case: Reliable message delivery
  • Benefits: Complex routing, delivery guarantees
  • Exchanges: Topic-based and direct routing

Event Store​

  • Use Case: Event sourcing and audit trails
  • Benefits: Complete event history, replay capability
  • Storage: Append-only event logs

Processing Patterns​

Publish-Subscribe​

  • Multiple consumers for single event
  • Decoupled service communication
  • Fan-out event distribution
  • Independent consumer scaling

Event Sourcing​

  • State derived from event history
  • Audit trail and compliance
  • Time-travel debugging
  • State reconstruction

CQRS (Command Query Responsibility Segregation)​

  • Separate read and write models
  • Optimized query performance
  • Command validation and processing
  • Eventual consistency handling

Saga Pattern​

  • Distributed transaction management
  • Compensating actions for failures
  • Long-running business processes
  • Cross-service coordination

Event Processing Guarantees​

At-Least-Once Delivery​

  • Events delivered at least once
  • Duplicate handling required
  • Higher reliability guarantee
  • Idempotency key usage

Exactly-Once Processing​

  • Deduplication mechanisms
  • Idempotency enforcement
  • Higher complexity but cleaner semantics
  • Critical for financial transactions

Message Ordering​

  • Per-partition ordering in Kafka
  • Global ordering where required
  • Timestamp-based sequencing
  • Causal consistency

Error Handling and Resilience​

Retry Mechanisms​

  • Exponential backoff strategies
  • Maximum retry limits
  • Dead letter queues
  • Circuit breaker patterns

Failure Recovery​

  • Event replay capabilities
  • State reconstruction
  • Partial failure handling
  • Compensation workflows

Monitoring and Alerting​

  • Event processing metrics
  • Lag monitoring
  • Error rate tracking
  • Performance dashboards

Integration Patterns​

Webhook Delivery​

  • HTTP-based event delivery
  • Retry and backoff logic
  • Signature verification
  • Filtering and routing

Stream Processing​

  • Real-time event analytics
  • Complex event processing
  • Windowing operations
  • Stream joins and aggregations

Batch Processing​

  • Periodic event processing
  • Bulk operations
  • Data warehouse integration
  • Historical analysis

Next Steps​