Due to cascading failures of computing centre air conditioning units starting around noon today, part of our computing resources had to be shut down.
This only affected farm nodes (including the clusters), APE, and some workgroup server like systems. All central services stayed up.
The latter systems are up again.
Cooling is back, with the exception of one device which may have been damaged and is still too hot to be serviced or tested.
For this reason, and because higher than average temperatures are expected for the next 24 hours, only 48 of the fastest and most energy efficient cores in the batch farm are currently enabled.
Operation will hopefully return to normal tomorrow.