Using the descheduler operator with the LowNodeUtilization strategy results in unstable/oscillatory behavior if the descheduler is used in this way to migrate VMs.  The problem appears to be due to the fact that when a VM is evicted, the eviction fails (returning a 429 error), but in fact the VM continues to be evicted in the background.  While this is happening, the descheduler assumes that the eviction has failed and keeps trying to evict every VM on the node (and anything else it needs to to get under the upper threshold).

Add a design proposal to support running KubeVirt in a cluster that also has a descheduler.