We became aware of a brief 10-minute disruption to GPU function scheduling on Google Cloud Platform. We have identified the cause of the incident as abusive traffic from a specific user, and have resolved the issue for now.

We are continuing to monitor.