Book Image

Learning RabbitMQ

By : Martin Toshev
Book Image

Learning RabbitMQ

By: Martin Toshev

Overview of this book

RabbitMQ is Open Source Message Queuing software based on the Advanced Message Queue Protocol Standard written in the Erlang Language. RabbitMQ is an ideal candidate for large-scale projects ranging from e-commerce and finance to Big Data and social networking because of its ease of use and high performance. Managing RabbitMQ in such a dynamic environment can be a challenging task that requires a good understanding not only of how to work properly with the message broker but also of its best practices and pitfalls. Learning RabbitMQ starts with a concise description of messaging solutions and patterns, then moves on to concrete practical scenarios for publishing and subscribing to the broker along with basic administration. This knowledge is further expanded by exploring how to establish clustering and high availability at the level of the message broker and how to integrate RabbitMQ with a number of technologies such as Spring, and enterprise service bus solutions such as MuleESB and WSO2. We will look at advanced topics such as performance tuning, secure messaging, and the internals of RabbitMQ. Finally we will work through case-studies so that we can see RabbitMQ in action and, if something goes wrong, we'll learn to resolve it in the Troubleshooting section.
Table of Contents (18 chapters)
Learning RabbitMQ
Credits
About the Author
About the Reviewers
www.PacktPub.com
Preface
Index

Performance tuning of RabbitMQ instances


Tuning the performance of a system is, in many cases, a nontrivial process that is conducted gradually over time. This also applies to the message broker itself. The RabbitMQ team has done a pretty good job in optimizing the various bits and pieces of the broker over time. One such example is topic exchanges. Version 2.4.0 significantly improved the performance of message routing from topic exchanges using a tire data structure. Another one is the significant improvement in performance predictability in version 2.8.1 during the heavy loading of the message broker due to improved memory management. However, there are many scenarios that require the tuning of the broker based on the usage patterns and properties of the system, as we shall see in this chapter.

To understand better how to tune the performance of our broker, let's take a look at the standard three-tier broker setup:

We can consider performance tuning at each level of message passing as follows...