The issue has been resolved. Between 4:53 PM and 8:53 PM, users experienced issues creating new compute instances and accessing newly created instances via SSH or the web console.

All compute instance operations are now working normally. Existing compute instances were not impacted at any point during this incident.

Compute metrics were temporarily unavailable between 2:10 PM and 5:05 PM. A fix has been deployed, and metrics processing has fully recovered. All systems are operating normally.