Outbox pattern for reliable events in Microservices - Architecture Diagram

System Overview - Outbox pattern for reliable events

The Outbox pattern ensures reliable event delivery in microservices by storing events in a local database table (the outbox) as part of the same transaction that updates the business data. A separate process reads these events and publishes them to the message broker, guaranteeing no events are lost even if failures occur.

This pattern helps maintain data consistency and reliable communication between services.

Architecture Diagram

User
  |
  v
API Gateway
  |
  v
Service A (Business Logic + Outbox DB)
  |
  v
Outbox Poller
  |
  v
Message Broker
  |
  v
Service B (Event Consumer)
  |
  v
Service B Database

Components

User

user

Initiates requests to the system

API Gateway

api_gateway

Receives user requests and routes them to Service A

Service A (Business Logic + Outbox DB)

service

Processes business logic and writes events to the outbox table in the same transaction

Outbox Poller

service

Reads events from the outbox table and publishes them to the message broker

Message Broker

message_queue

Delivers events asynchronously to interested services

Service B (Event Consumer)

service

Consumes events from the message broker and updates its own database

Service B Database

database

Stores data updated by Service B based on received events

Request Flow - 7 Hops

User → API Gateway

API Gateway → Service A (Business Logic + Outbox DB)

Service A (Business Logic + Outbox DB) → Service A (Business Logic + Outbox DB)

Outbox Poller → Service A (Business Logic + Outbox DB)

Outbox Poller → Message Broker

Message Broker → Service B (Event Consumer)

Service B (Event Consumer) → Service B Database

Failure Scenario

Component Fails:Message Broker

Impact:Events cannot be delivered to Service B, causing delays in event processing

Mitigation:Outbox Poller keeps events in the outbox table until broker is available; retries publishing events; message broker should be highly available with replication

Architecture Quiz - 3 Questions

Why does Service A write events to the outbox table in the same transaction as business data updates?

A To ensure events are not lost if the business update fails

B To speed up event delivery to Service B

C To avoid using a message broker

D To reduce database storage

Design Principle

The Outbox pattern ensures reliable event delivery by combining business data updates and event recording in a single transaction, then asynchronously publishing events. This avoids lost or duplicated events and maintains data consistency across microservices.