Due to how the 2 hour delay is set via an annotation on the ignition token, if the reconcile loop runs more than every 2 hours, the timestamp is reset and the expiration date is never reached, thus cleanup does not occur.

    1.Deploy a cluster with Hypershift Operator
    2.Update the nodepool(s) to cause a new ignition token(s) to be generate
    3.When the reconcile loop in HO runs more often than every 2 hours, watch the expiration timestamp annotation.
    

    Timestamp is reset instead of allowing the initial 2 hour window to expire

    Timestamp is not reset so that the 2 hour window set by the timestamp can expire and the ignition-server performs the cleanup