New to KubeDB? Please start here.

Druid Compute Resource Autoscaling

This guide will give an overview on how KubeDB Autoscaler operator autoscales the database compute resources i.e. cpu and memory using druidautoscaler crd.

Before You Begin

• You should be familiar with the following KubeDB concepts:
  • Druid
  • DruidAutoscaler
  • DruidOpsRequest

How Compute Autoscaling Works

The following diagram shows how KubeDB Autoscaler operator autoscales the resources of Druid database components. Open the image in a new tab to see the enlarged version.

  

The Auto Scaling process consists of the following steps:

1. At first, a user creates a Druid Custom Resource Object (CRO).

2. KubeDB Provisioner operator watches the Druid CRO.

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

4. Then, in order to set up autoscaling of the various components (ie. Coordinators, Overlords, Historicals, MiddleManagers, Brokers, Routers) of the Druid cluster the user creates a DruidAutoscaler CRO with desired configuration.

5. KubeDB Autoscaler operator watches the DruidAutoscaler CRO.

6. KubeDB Autoscaler operator generates recommendation using the modified version of kubernetes official recommender for different components of the database, as specified in the DruidAutoscaler CRO.

7. If the generated recommendation doesn’t match the current resources of the database, then KubeDB Autoscaler operator creates a DruidOpsRequest CRO to scale the database to match the recommendation generated.

8. KubeDB Ops-manager operator watches the DruidOpsRequest CRO.

9. Then the KubeDB Ops-manager operator will scale the database component vertically as specified on the DruidOpsRequest CRO.

In the next docs, we are going to show a step-by-step guide on Autoscaling of various Druid database components using DruidAutoscaler CRD.