Services that publish and consume messages form the core of event-driven systems. Testing them thoroughly ensures that important events are emitted correctly and that consumers handle messages safely, including malformed or duplicate ones. This is fundamental for reliable business workflows.
Testing Message Producers
Producer tests verify that services publish events with correct schemas, routing keys, and headers when important actions occur. You can use test doubles for the messaging broker or capture messages from a dedicated test topic. Assertions focus on event structure, content, and frequency.
# Example producer test ideas
- Verify an OrderCreated event is published when an order is placed.
- Check that event payload includes orderId, userId, and total.
- Ensure sensitive data is not leaked in event fields.
- Confirm events are not duplicated on retries (if required).Consumer tests verify that services process incoming messages correctly, handle errors, and interact with other components like databases and APIs as expected. They should also cover how consumers behave when messages are missing fields, duplicated, or delivered in unexpected orders.
Testing Message Consumers
Approaches include driving consumers with test messages on dedicated queues, using in-memory brokers for unit-like tests, or running integration tests against real brokers. Assertions might check database state, emitted follow-up events, or calls to dependent services.
Common Mistakes
Mistake 1 — Testing only producers or only consumers
Both sides need coverage.
❌ Wrong: Verifying emitted events but never testing how they are consumed (or vice versa).
✅ Correct: Design tests for both production and consumption behaviours.
Mistake 2 — Ignoring error paths in consumers
Bad messages are inevitable.
❌ Wrong: Assuming all messages are well-formed and processed successfully.
✅ Correct: Test how consumers handle malformed, duplicate, or failing messages.