RabbitMQ, Kafka, and Redis: Which Broker to choose for a MicroService?

Asynchronous communication is typically managed through a message broker. There are other ways also, like Async IO, but they’re more scarce and limited.

women and man sitting in-front of monitor to make decission
Photo by Lagos Techie on Unsplash

When picking a broker for executing your asynchronous operations, you ought to consider a couple of things:

  1. Broker Scale — the number of messages sent per second within the system.
  2. Consumer efficiency — whether the broker is efficient in managing one-to-one and/or one to many consumers.
  3. Data Persistency — the power to recover messages.



  • RabbitMQ has some SaaS-managed services, but it’s not part of the native major cloud provider stack.
  • It supports all major languages, including Python, Java, .NET, PHP, Ruby, JavaScript, Go, Swift, and more.
  • Performance lag is expected when in persistent mode.


  • Kafka has managed SaaS on top native cloud vendors like Azure, AWS, and Confluent.
  • Kafka supports all major languages, like .NET, C/C++, Python, Java, PHP, Ruby, JavaScript, Go, Swift, Clojure, and more.


  • It is a bit different from the other message brokers, from the way it is stored and processed.
  • At its core, it is an in-memory data store that can be used as a highly performing app, but for transient uses, it can be used as a message broker.
  • Redis dumps its memory into a Disk/DB but well perfect for real-time data processing.

Use case:

  • For short-lived messages, prefer Redis.
  • While if your application has large amounts of data, then Kafka would be a better option.
  • If the data is complex enough, then RabbitMQ will do its job.

Message brokers are becoming increasingly important as companies upgrade their applications in readiness for the cloud. Many of the world’s most successful firms rely on Payoda’s message broker capabilities, which are designed to support today’s agile development environments.

Let’s Talk!

Authored by: Starlin Daniel Raj