So You Want to Maintain a Reliable Event-Driven System
summary
Advice for running event systems: publish explicit versions (for example use order-confirmed-V1 topics) and set a hard deprecation date; use the outbox pattern to avoid lost state/events and spin up a test harness that subscribes to your bus to assert events actually arrive; adopt the cloud events specification and include a W3C traceparent so OpenTelemetry can link traces; validate emitted JSON with njschema/JSON Schema, retry tests to SLOs (e.g. 200 ms) and inspect the Plant‑Based Pizza GitHub for a runnable example.
AI-written summary · about
chapters (10)
- 0:00Event-driven architecture
- 2:55API first design
- 7:38Testing asynchronous systems
- 14:21Pattern based testing
- 20:56Testing sad paths
- 25:06Contract testing
- 27:47Observability and tracing
- 36:41Span linking vs propagation
- 40:03Event first design
- 45:26Summary and takeaways
related talks
55mSo You Want to Maintain a Reliable Event Driven System
54mSo You Want to Build An Event Driven System?
52mSo You Want to Build An Event Driven System?
52mSo You Want to Build An Event Driven System?
31mThe Power of Event-Driven Systems without Burning your Hands or Budgets
55mAvoiding mistakes with events, one event at a time
30mUnhappy Path & Dealing with Bad Events
49mAvoiding mistakes with events, one at a time