You are looking at the documentation of a prior release. To read the documentation of the latest release, please visit here.

New to KubeDB? Please start here.

RabbitMQ Horizontal Scaling

This guide will give an overview on how KubeDB Ops-manager operator scales up or down RabbitMQ cluster.

Before You Begin

How Horizontal Scaling Process Works

The following diagram shows how KubeDB Ops-manager operator scales up or down RabbitMQ database components. Open the image in a new tab to see the enlarged version.

The Horizontal scaling process consists of the following steps:

  1. At first, a user creates a RabbitMQ Custom Resource (CR).

  2. KubeDB Provisioner operator watches the RabbitMQ CR.

  3. When the operator finds a RabbitMQ CR, it creates required number of PetSets and related necessary stuff like secrets, services, etc.

  4. Then, in order to scale the RabbitMQ cluster, the user creates a RabbitMQOpsRequest CR with desired information.

  5. KubeDB Ops-manager operator watches the RabbitMQOpsRequest CR.

  6. When it finds a RabbitMQOpsRequest CR, it halts the RabbitMQ object which is referred from the RabbitMQOpsRequest. So, the KubeDB Provisioner operator doesn’t perform any operations on the RabbitMQ object during the horizontal scaling process.

  7. Then the KubeDB Ops-manager operator will scale the related PetSet Pods to reach the expected number of replicas defined in the RabbitMQOpsRequest CR.

  8. After the successfully scaling the replicas of the related PetSet Pods, the KubeDB Ops-manager operator updates the number of replicas in the RabbitMQ object to reflect the updated state.

  9. After the successful scaling of the RabbitMQ replicas, the KubeDB Ops-manager operator resumes the RabbitMQ object so that the KubeDB Provisioner operator resumes its usual operations.

In the next docs, we are going to show a step by step guide on horizontal scaling of RabbitMQ database using RabbitMQOpsRequest CRD.