Delayed scaling due to node failure

Updates

• resolved

  All affected models have been scaling correctly for more than 30 minutes at this point, and we see no residual prediction queues. Thank you for your patience!

  Aug 10, 2026, 9:45 PM UTC

• monitoring

  Scaling decisions were delayed by nearly 1 hour after the controller responsible for emitting queue metrics failed to schedule on a soft-failed node. We have since cordoned and drained the node, and the controller is emitting queue metrics again.

  Aug 10, 2026, 9:05 PM UTC