This incident has been resolved. No user traffic was affected - all requests were handled by our stable service

During a routine deployment, an internal configuration change introduced an inconsistency in our traffic routing system, causing an internal reconciliation loop. No user traffic was affected - all requests were served by the stable service throughout. The affected deployment was rolled back and normal operation was restored.

We are currently investigating an issue affecting the API service. Our team has identified the root cause and is actively applying a fix. No data loss is involved. Some API requests may experience degraded performance during this time.
ETA for resolution: ~15 minutes.
We will update this page once the issue is resolved.