Storage volume attachment degraded

Our storage orchestration service (which coordinates volume attachments) failed to restart due to an internal metadata inconsistency, leaving it temporarily unable to serve volume attach/detach requests.

Running workloads kept their already-attached storage and were unaffected. The inconsistency was corrected and the service restored.

We are currently addressing an issue with our storage control plane. As a result, attaching storage volumes is temporarily failing.

What is impacted: starting or restarting a virtual machine, and creating new instances that require storage. These operations may hang or fail during this incident.

What is NOT impacted: virtual machines and workloads that are already running continue to operate normally, as their storage is already attached. No data is at risk.

Our team has identified the cause and is working on restoring normal operations. We will update this page as the situation progresses.