zoom_in Click to enlarge

RabbitMQ

▲ ▼

language

description RabbitMQ Overview

Messages pass between applications through RabbitMQ, an open-source broker used to separate the system producing work from the system consuming it. Producers publish messages, and exchanges route them to queues according to configured rules. Consumers then receive and process the queued messages. This routing model is useful when several services need different subsets of the same stream of work.

RabbitMQ supports AMQP and provides acknowledgments and durability options that can help applications handle failures. It commonly supplies the messaging layer for distributed task systems such as Celery, while leaving task execution and orchestration to those systems. It suits teams that need explicit routing and control over message handling across servers. Delivery behavior depends on configuration and application logic, so consumers may need to handle repeated messages safely rather than assuming each task can run only once.

help RabbitMQ FAQ

Does RabbitMQ restart failed worker processes by itself?

RabbitMQ is a message broker, not a process supervisor. It can keep durable queues and messages available, but tools such as systemd, Kubernetes, Supervisor, or Celery worker management handle restarting crashed worker processes.

Why do Celery deployments often use RabbitMQ?

Celery supports RabbitMQ as a broker because RabbitMQ implements AMQP 0-9-1 and handles task delivery, acknowledgements, and routing well. Redis is also common, but RabbitMQ is often chosen when routing and delivery guarantees matter.

What are RabbitMQ exchanges used for?

RabbitMQ sends messages through exchanges before they reach queues. Common exchange types include direct, topic, fanout, and headers, which let one producer route work to many different queues without hard-coding every consumer.

How does RabbitMQ help after a worker crashes mid-task?

If the consumer uses manual acknowledgements and the message is not acked before the crash, RabbitMQ can requeue it for another worker. For persistence across broker restarts, the queue and messages also need to be declared durable or persistent.

swap_horiz

Looking for RabbitMQ alternatives? Compare top competitors ranked & scored

arrow_forward