WHAT

Investigate the changes required to change the data plane worker node type for our standard data plate fleet. The outcome of this JIRA should just outline what needs to change and the level of effort required, no actual changes should be progressed since the Kafka Fleet Services team may determine the current node type is the most optimal.

WHY

The memory optimised (R class) ec2 instances may be more cost-effective than the current general purpose (M class) ec2 instances we use for the worker nodes that the RHOSAK instances are deployed to.

HOW

This will likely require a new machine pool to be created, then scaling down the existing machine pool to allow the k8s scheduler to move the pods.

DONE

• Document changes required in kas-fleet-manager
• Estimate level of effort