This incident has been resolved.

We are monitoring the issue. We applied a firmware change that we think may prevent this from happening again. We will monitor this issue over the next few days since this has happened twice now.

A cloud hypervisor has rebooted, we are investigating.