How do I track down random reboots on Server 2019?

Check the System event log around each reboot first. Look for a bugcheck, an unexpected shutdown, and any errors just before it. An unexpected shutdown event tells you the server did not shut down cleanly, but not necessarily why.

If there is a dump, analyze it and compare the results across crashes. One reported case pointed repeatedly at libusb0.sys while Eaton Intelligent Power Protector was installed, but that alone does not prove the UPS software caused the reboots. Check which package installed the driver and raise the findings with the vendor.

Since dumps are missing sometimes, check that crash dump settings are enabled and that the system has enough disk space and a suitably configured page file. A sudden loss of power or a reset can prevent a dump from being written, so no dump does not rule out hardware or power trouble.

Build a timeline from the server logs and any centralized monitoring you have. Compare it with UPS events and hardware management logs, if available. If the evidence points toward the UPS connection, discuss a controlled test or alternate setup with the vendor. Plan it carefully so the server stays protected from power loss.

Don't get too attached to the UPS theory yet. If this is a hypervisor, check other low-level software too, including security products and their drivers. Compare versions and crash evidence before changing several things at once, or you may never know what helped.