Communities experienced approximately 2 minutes of degraded availability starting at 10:26 AM EDT. The issue was caused by application servers returning 500 errors due to a combination of unoptimized database queries and a framework regression that affected health checks, causing pods to cycle. Multiple engineering teams responded and deployed a fix.
Root cause: A Rails 8.1 upgrade changed how internal state was managed between requests, causing intermittent crashes. This was compounded by high-frequency queries that increased database load. Both issues have been patched.