I found a possible correlation between migrations and the cpu problem so I'll definately check this out the next time a VM peaks.
The biggest problems were with 2 critical servers that we changed to manual automation level to prevent them migrating and the problem hasn't occurred on these servers since then. This strengthens the likeliness of this being related to the kb article you posted.
I'll give you the points for a correct answer once I verify. Thank you very much so far!
Chris
