Glossary

Message broker

Middleware that routes messages between producing and consuming systems so they don't need to connect to each other directly.

A message broker is software that sits between systems that produce messages and systems that consume them, accepting a message from a sender and routing it to one or more recipients. Producers and consumers connect only to the broker, not to each other, so either side can be added, removed or restarted without the other noticing.

Traditional brokers like RabbitMQ and ActiveMQ follow a queue model: a message is typically delivered to one consumer and then removed. This differs from event streaming platforms such as Apache Kafka, which keep messages on a durable, replayable log that multiple independent consumers can read at their own pace. Both are "message brokers" in the general sense, but the queue model suits task distribution and request routing, while the log model suits capturing a history of events.

Message brokers matter because they decouple systems in an event-driven architecture: a producer doesn't need to know who is listening, and a slow or temporarily offline consumer doesn't block the producer. Common pitfalls include using a broker as a long-term datastore it was never designed to be, letting queues grow unbounded when a consumer falls behind, and underestimating the operational work of running a broker reliably at scale.

Last reviewed September 22, 2026

In the index now

Related terms

Apache Kafka An open-source distributed event streaming platform that stores records in ordered, durable, replayable logs called topics. Event streaming Continuously publishing and consuming records of things that happened, in order, through a durable, replayable log. Event-driven architecture (EDA) A system design where components communicate by producing and reacting to events, rather than calling each other directly. Webhook An automated HTTP callback a system sends to a URL the moment a specific event happens, instead of waiting to be asked.