Consumer Patterns

Consumers determine whether events become clear downstream capabilities or a chain of brittle hidden dependencies. Consumer design affects throughput, idempotency pressure, local state management, and how easy the platform is to operate when something fails. Chapter 6 focuses on the main consumer-side patterns directly so those responsibilities are explicit rather than treated as “whatever reads the message.”

Read the lessons in order. The first covers stateless consumers as the simplest consume-process-acknowledge model. The second explores stateful consumers and projections, where replay and correctness become more demanding. The third explains competing consumers as a throughput pattern with ordering trade-offs. The fourth covers enrichment, translation, and routing consumers, especially where those patterns become useful and where they become opaque integration glue.

If your event-driven platform has many consumers but nobody can clearly explain which ones are simple handlers, which ones maintain state, and which ones are just reshaping messages for the next hop, Chapter 6 is the place to clean up that consumer model.

In this section

• Stateless Consumers
  A practical lesson on stateless consumers as simple consume-process-acknowledge handlers, including why they scale well and where idempotency still matters.
• Stateful Consumers and Projections
  A practical lesson on consumers that maintain read models, workflow state, or derived local views, including why replay and correctness become more demanding.
• Competing Consumers
  A practical lesson on competing consumers as a throughput pattern, including partitioning, hot keys, and the trade-off between scalability and ordering guarantees.
• Enrichment, Translation, and Routing Consumers
  A practical lesson on consumers that enrich data, translate schemas, or route events for downstream domains, including where they add value and where they become opaque glue.