Between 14:25 UTC on 28 October and 11:46 UTC on 29 October, creating or cloning an instance in the Portal failed with an error for accounts hosted in our West Europe region. Existing instances were not affected.

The cause was that the database capacity we keep for new instances in West Europe had run out. We added capacity at 11:45 UTC on 29 October, and instance creation has worked normally since. If an instance could not be created during this window, please try again. We are also improving our capacity monitoring so that it warns us well before this capacity runs out.

Root cause: a power failure at a datacenter operated by our cloud provider in West Europe. When utility power was interrupted, several backup generators did not start, and part of the provider's infrastructure lost power, including the database infrastructure behind a subset of hosted instances. Service returned at 16:11 UTC as power was restored. We have since added monitoring that alerts us within minutes if a group of hosted databases becomes unreachable.