The datacenter that we run on just informed me that they had a storm go through and are experiencing overheating of equipment due to loss of cooling. So we may have poor performance or go offline today. If necessary I will keep you updated here:
It’s also definitely possible for a storm to damage the air conditioning systems whether they still have power or not, which is exactly what sounds like happened. They didn’t give more details, but I have no reason to doubt them:
Specific servers and individual network hardware does not necessarily run more slowly under overheating conditions (although there are definitely some nuances there with CPU scaling), but virtual services comprised of multiple nodes can run slower if they have to take down some nodes to reduce the power/heat load. Also if they are running multiple peering networks some of them might be offline from the storm and/or they might have to shut down some of the routers to reduce power/heat.
At any rate, it looks like they got things back to normal according to the most recent update.