Top 10 Event-Driven Architecture Pitfalls
summary
Ten concrete failure modes from real workshops at 150 companies: duplicate deliveries, out‑of‑order processing, race conditions, test flakiness from Spring test context and concurrent consumers, lost messages, dual‑write/transaction ordering mistakes, and insecure state handling. Examples use Kafka partitioning to preserve order, Spring Boot concurrency pitfalls, optimistic/pessimistic locking failures, event‑carried state vs HTTP GET races, the outbox pattern as the fix for atomic send+save, and a real “cheese” infinite‑loop event tsunami that crashed production.
AI-written summary · about
chapters (11)
- 0:00Introduction to EDA pitfalls
- 0:51Common consumer challenges
- 5:32Testing event-driven systems
- 10:18Managing race conditions
- 14:28Partitioning and message ordering
- 18:14Advanced consumer patterns
- 24:26Consumer callback race risks
- 30:20Producer and consistency issues
- 33:26Dual write and atomicity
- 39:08Idempotency and message types
- 43:15Conclusion and final summary
related talks
47mTop 10 Event-Driven Architecture Pitfalls
29m9 Fallacies of Event Driven Architectures
42mApplication Architectures: The shadows and highlights of Event Driven Design
51mComplex Event Flows in Distributed Systems
26mTen Best Practices for Event Streaming Success
53mPatterns of Event Driven Architecture
54mReacting to an Event-Driven World
47mComplex event flows in distributed systems