The machine ran fine for months and now reboots by itself a few times a day at no fixed time. Where do we start?

Work through this order, most common cause first: 1. Power. Measure the voltage at the machine's power terminals while the line runs at full load, ideally with a meter that records minimums. Contactors or drives starting up can cause a brief dip that reboots the machine, while a measurement taken at idle looks fine. Check for loose ferrules too. 2. Heat. Read CPU temperature in BIOS or a monitoring tool and feel the heatsink. A fanless machine running for months in a dusty workshop often has a thick layer between the fins; clean it and watch again. 3. System logs. In Windows, Event Viewer filtered on Kernel-Power (event 41) distinguishes a sudden power loss from a software fault; WHEA entries point to hardware. On Linux, read journalctl for the previous boot. 4. RAM and storage. Run an overnight memory test and check the SSD's S.M.A.R.T. data. 5. Watchdog. Many industrial machines have one; if your application enables the watchdog and a thread hangs, you get exactly this reboot pattern. If it is still unclear, send us the model number and log screenshots and we will look at it with you.