Skip to main content

Investigation RoadMap

Specs of Instances​

...WIP...

This roadmap compares single-node and clustered RabbitMQ behavior. Here, “churn” means repeatedly creating and deleting a resource under load.

RabbitMQ terminology: a connection is a network session; an AMQP channel is a lightweight lane inside it; a queue holds messages for consumers; an exchange routes published messages; and a binding is the routing rule from an exchange to a queue. The Event Exchange publishes broker lifecycle events for observation.

For both Cluster and Single-node, answer the following:​

  1. What's the impact and what can you observe out of:
    1. high churn of Connection?
    2. high churn of AMQP channels?
    3. high churn of Exchanges?
    4. high churn of Queues?
    5. high churn of Bindings?
  2. What can you observe if you enable Event Exchange?
    1. For each type of binding
    2. In various types of loads and scenarios
  3. What are the best use cases for RabbitMQ federation? What can it do across independent brokers?
  4. What are the limitations of quorum queues? How do they compare with Classic queues?
Share: