Between August 8 and 10, some Buffer customers experienced intermittent sign-outs, failed post publishing, and issues loading parts of the product. We identified 23,977 posts that failed during the affected period, across several social networks, and received 52 related customer-support conversations. The issue has been resolved, and monitoring confirmed the symptoms did not recur.

The underlying cause was an intermittent DNS failure affecting a small number of AWS infrastructure hosts. This disrupted requests to both Buffer services and external platforms; our existing authentication behavior could also interpret a temporary network failure as an invalid session, causing repeated logouts. We removed the affected infrastructure from service, added safeguards so temporary connectivity failures preserve customer sessions, and are adding earlier alerting for DNS resolver errors.