Publish/subscribe is often the first pattern teams adopt when moving toward asynchronous, message-based architectures. It’s a powerful way to decouple services and enable extensibility—but on its own, pub/sub doesn’t solve many of the problems that show up in real distributed systems, such as data loss, broken workflows, or systems that struggle under load.

In this session, we’ll go beyond basic pub/sub to explore a set of proven messaging patterns used in production. We’ll look at how the Outbox pattern prevents message loss during failures, how sagas coordinate long-running workflows without distributed transactions, and how competing consumers enable elastic throughput when demand spikes. Along the way, we’ll also examine essential operational safeguards such as retries, dead-letter queues, back pressure, and idempotency.

By the end of the talk, you’ll understand when and how to apply these patterns to build systems that are not just decoupled, but resilient, scalable, and designed to handle failure gracefully from day one.