Pattern: Event-driven architecture with message broker

Services are tightly coupled through synchronous calls. Adding a new consumer requires modifying the producer. Failures cascade across services.

Use events and a message broker for loose coupling:

# Producer: publishes events without knowing consumers
class OrderService:
def __init__(self, db, event_bus):
self.db = db
self.event_bus = event_bus

def place_order(self, order_data):
order = self.db.save(order_data)

# Publish event - don't care who listens
self.event_bus.publish('order.placed', {
'order_id': order.id,
'user_id': order.user_id,
'total': order.total,
'items': order.items,
})
return order

# Consumer 1: Email service
@event_bus.subscribe('order.placed')
async def send_confirmation(event):
await email.send(event['user_id'], 'Your order is confirmed')

# Consumer 2: Analytics (added later, no change to producer!)
@event_bus.subscribe('order.placed')
async def track_analytics(event):
await analytics.track('purchase', event['total'])

# Consumer 3: Inventory
@event_bus.subscribe('order.placed')
async def reserve_inventory(event):
for item in event['items']:
await inventory.reserve(item['sku'], item['qty'])

Benefits:

• Add consumers without touching producers
• Services can fail independently
• Natural audit log (event stream)
• Replay events for debugging or rebuilding state

Tools: RabbitMQ, Kafka, Redis Streams, AWS SQS/SNS, NATS

Events decouple producers from consumers. New features are added by subscribing to events, not modifying existing code.