From 6:00pm CET, we observed underperformance of some assets on our primary cloud provider. Affected assets were migrated over the course of 15 mins. During this time, performance was substantially degraded for some users.
The issue was identified and resolved. The root cause was found to be that regular maintenance actions on a small table in the cloud database blocked transactions. These actions have been shifted to a lesser-used read-only replicated copy.