I think you are right one this.
You could also look at using storage DRS with your profiles, and that way if a datastore is getting crunched and the otehr datastores are not as busy the workload could move and help keep service at the levels you require.