CircleCI Outage Traced to Scheduled Workflow Deletions Its Earlier Fix Missed
CircleCI's pipeline and workflow processing stopped for about 12 hours on October 9, 2026, after its Postgres database ran out of disk space. The trigger was a scheduled cleanup of completed Temporal workflow history, not an unexpected load spike. A fix shipped after a near-identical October 6 outage had moved the transaction log to a dedicated volume and reduced write pressure, but it did not account for deletions already scheduled for October 9. Once the log volume filled, Postgres went read-only, and a later failure from bloated cleanup tables and a retry storm caused further slowdowns before the backlog cleared.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.


Discussion (0)
Log in to join the discussion and vote.
Log in