There are many causes of unknown server reboots, but in my case it was issues with power strips and incorrect connections to the UPS (Uninterruptible Power Supply).
I re-routed the power cables. Previously, redundant power was lost on some servers and the offending connections were being made at a power strip which daisy-chained to the UPS. The servers were temporarily connected to the UPS via smaller cables and the offending ones disconnected; either the cables were bad or the power strip was bad. I believed the power strip was bad, because connecting via a strip is incorrect as the UPS is more reliable. One really old power strip was removed completely, the old, large cables were reconnected, with most of the server power supplies being connected to the UPS. One power strip was retained and one redundant connection from each of two non-critical servers was connected here. Not a perfect solution, but it works.
