Hypothetically, if a cluster is setup with some sort of node pool provider (VSphere), and a cluster change was initiated through Rancher (Update 1.28.x to 1.29.x). After which Rancher notices that VM config doesn't line up with expectations on vCenter (due to things having been moved around). Is there a single deployment or pod that I can halt, to stop the deletion and creation of new VMs until I can figure out how to rollback? What is actually responsible for creating new Infrastructure to reconcile the state of a node pool to the true state?