We have a webrole running our application with two instances. One of the instances in one of our webroles just unexpectedly stopped. Initially, we thought it might be due to a role instance upgrade but it was down for more than an hour. It came up after an hour and went down again. We scaled up to three instances to see if the 3rd instance works ok. We found that instance 1 and 3 works fine but still the instance number 2 stops. Again, the other two instances seem fine. Does this mean that the VM instance is somehow corrupted?