One hypervisor only responds to ping. It does not take new VMs anymore and does not delete VMs that should be deleted.
19:57 UTC: We are going to reboot it. Some databases (that run on this hypervisor) will become unresponsive for a few minutes.
20:18 UTC: Hypervisor has been rebooted. All services hosted on it have been checked: everything is up and running.
Logs show a kernel panic.