The affected node is back online and virtual machines are now powering back on. Services should return to normal over the next 10 minutes as each VM completes boot.

We are currently performing an emergency replacement of a failed NVMe drive from one of the RAID10 mirrors. All customer data is safe and intact. Affected VMs will be back online as soon as the disk replacement is complete and the node has been booted back into the OS. We'll post an update once service is restored.