We've identified the cause and paused a problematic background query. Services have recovered, we're working on a proper fix for the query.

The database for insights is currently unstable because of a surge in RAM usage. Cause is currently being investigated.