OpenShift Virtualization - Cluster Sizing Guide

This document provides guidance, via an example, of how to size an OpenShift cluster to host virtual machine-based workload. The values used in this document represent reasonable values, however it is important to use, as closely as possible, virtual machine sizes that reflect the real expected workload of the cluster. In the absence of known virtual machine sizing data, using the example sizes in this document will suffice. Regardless of the sizing data input and the result of the exercise, it’s important to use the monitoring and capacity management tools available to continue to verify and validate all assumptions made, with follow up actions to adjust cluster capacity as needed.

The steps described and used in this document for sizing include:

• Determining the raw capacity needed for the workload
• Factoring overcommitment into the capacity needed
• Identifying node size and the impact to the cluster
• Adjusting node count for cluster architecture
• Further adjusting capacity requirements to account for high availability
• Considerations for special resources