Document link: https://docs.openshift.com/container-platform/4.16/machine_configuration/machine-config-node-disruption.html

There are many node disruption policy examples that instruct to reload chronyd.service after a change. The problem is that chronyd.service doesn't admit reloading, so applying any of the chronyd examples results in immediate failure and the node to become degraded for the MCO.

The examples must be replaced with restarting the service.