Current task contains several steps:

1) The new descheduler profile "DevKubeVirtRelieveAndMigrate" requires the kernel argument "psi=1" on worker nodes. It may take tens of minutes to set this argument before running tests and remove during the teardown, so we decided to add it during the cluster installation

2) Moving kube-descheduler operator installation part to the cluster installatiion level also will help to significantly reduce the test run time

3) Need to refactor existing descheduler automation

4) Add automation for new tests 

5) update CI tier3 scheduled job - set the flag for descheduler installation