Previous incidents

Earlier today, Sift experienced a database storage incident caused by a deliberately abusive user behaviour. We identified and suspended the abusive account, stopped the runaway import impact, and began removing the incident-related data. The system is now recovering while we implement hardening.

We identified overlapping feed refresh workers as the cause of the backlog and have deployed safeguards to clear the queue and reprocess feeds gradually while protecting database capacity. This should be resolved by 20:00 BST.