The basic operation of OpenShift workload scheduling does not take into account some specific characteristics of persistent workloads like VMs. Today customers reply on vSphere Distributed Resource Scheduler to balance workloads on their clusters over time, across infrastructure life cylce events: upgrades, expansion, failures and recovery.
These components should be more aware of long-running workloads, and consider actual CPU and memory consumption, not just what was requested. As these platform capabilities evolve, they needs to act appropriately with VM workloads.

Goal: Ensure that VMs integrate seamlessly into the load aware scheduling and descheduling features of OpenShift.