What is the difference between a message queue (like RabbitMQ) and a job queue (like BullMQ), and when would you choose one over the other?

A job queue is typically purpose-built around the concept of discrete, trackable units of work with retries, progress, and completion status, often backed by Redis, and tightly integrated into a single application's worker model. A message queue like RabbitMQ (or Kafka) is a more general-purpose messaging infrastructure supporting complex routing topologies (exchanges, multiple consumer groups, fanout patterns), designed for inter-service communication across a distributed system rather than just offloading a single application's background work -- appropriate when multiple independent services need to publish and subscribe to events, not just one application processing its own job backlog.

// Job queue (BullMQ): task-oriented, within one application
await imageQueue.add('resize', { imageId, sizes: [100, 500] });

// Message queue (RabbitMQ): event-oriented, across services
channel.publish('orders_exchange', 'order.created', Buffer.from(JSON.stringify(order)));
// Multiple independent services (inventory, shipping, analytics) can each
// subscribe to 'order.created' events without the order service knowing about them