This message was deleted.

By chance do you have persistent logging redirected?

We actually have a sort of similar issue. Multi-OS, reboot cycle script - every now and then, after the script has done it's thing and the VDA is online again, another reboot or even power off occurs. But this is within minutes, so no great risk to users aside from annoyance. So keeping 👀 on this thread. 🙂

check event logs on the delivery controllers as well, and if you have 'logging' enabled in studio, check there. LIke @Jon Bucud mentioned, hopefully you have the event logs on the VDA redirected to a persistent location.