This is pretty neat. We accomplished something similar but simpler. We use unattended reboots with randomised selection of time throughout 24 hours. Drain node

This is pretty neat. We accomplished something similar but simpler. We use unattended reboots with randomised selection of time throughout 24 hours. Drain node with a script. Not perfect but simple to do and setup.