How to Read Windows Event Viewer Logs Well Enough to Find Out Why a PC Rebooted Itself
Two machines this month came to me with the same story: it rebooted by itself overnight and nobody touched it. One was a reception PC, the other a laptop that lives on a dock. Both owners were sure it was a virus. Neither was.
Windows keeps a very detailed diary of everything that happens on the machine, and it has been doing it the whole time. The trouble is that the diary is written for engineers and it opens on a page full of noise. If you learn which three or four entries to look at and ignore the rest, you can usually answer the reboot question in under ten minutes. Here is how I actually do it, without pretending you need to read the whole thing.
Open the viewer and get to the right log first
Event Viewer is a stock Windows utility. Nothing to install. Hit Start, type Event Viewer, open it. On the left you get a tree. Expand Windows Logs and click System. That is the one you want. Not Application, not Security. The System log is where Windows records its own housekeeping, and that includes shutdown and startup reasons.
Now the part that saves you time: clear the clutter before you read. On the right pane, click Filter Current Log. In the Event level box, uncheck everything except Critical, Error, and Warning. That alone cuts the list down by a huge margin. If you are chasing a surprise reboot, you can also type 41 into the Event IDs field, or a few comma-separated ones, and see only those. You can always remove the filter later, and doing so is fine.
The four IDs that answer most reboot questions
Windows stamps specific event IDs onto every unexpected shutdown. You do not need to memorise the whole catalog. Learn these and you are most of the way there.
- 1074 in System. Something asked the machine to restart, and Windows logged what. The event usually names the process and a reason code. If you see a Windows Update process here, your reboot is explained. If it is an app or a scheduled task, that is your culprit.
- 6006 and 6005. A clean shutdown and a clean startup. If you see 6006 right before the unexpected restart, the machine shut down politely, which points at software or a scheduled job. If you do not, look at the next pair.
- 6008. “The previous system shutdown was unexpected.” This is the one you see when the machine crashed, lost power, or was hard-reset. It tells you it happened but not why. Treat it as a bookmark.
- 41. The famous Kernel-Power event. The system rebooted without cleanly shutting down first. I have to be honest with you about 41: it is a symptom, not a diagnosis. It shows up after a power cut, a thermal trip, a hang, or a hardware fault. On its own it tells you the machine died, and not much more. Newer builds sometimes attach extra detail, a bugcheck code or a bit indicating the power button was held, so read the whole entry rather than just the summary line.
Work the sequence in order. If 1074 fired, you have a named process and you are basically done. If you get 6008 plus 41 with no 6006 in front, the machine did not shut down on purpose. That is the moment to stop reading logs and go check hardware.
Where to look when 41 has no obvious cause
A Kernel-Power 41 with nothing else around it sends me in three directions.
Blue screens first. A crash bugcheck leaves its own entry, usually ID 1001 with source BugCheck, sometimes with a dump code. Match the code to a known issue and you have your answer. If you see repeated 1001s, an outdated driver is the usual suspect. That is a separate problem from a clean power loss and worth fixing on its own schedule.
Power and temperature. A failing PSU, a swollen laptop battery, or a fan clogged with dust will all produce 41s. Check for thermal events in the same window. If the reboots cluster at a certain time of day or after the machine has been busy, temperature is a good guess. If your office has flaky mains, a power blip shows as an unexpected shutdown and nothing else.
Scheduled maintenance. Windows Update, driver updates, and some third-party app updaters all trigger restarts through 1074. Filter for 1074 alone and read the reason. If it is an update, make peace with it or set active hours. If it is a vendor tool you do not need, uninstall it.
The habit that actually saves you time
Whenever someone reports a mystery reboot, my first move is to freeze the evidence before the machine runs another twenty minutes. Right-click System in the tree, choose Save All Events As, and drop the file onto your own drive. Later reboots get written on top of the old noise and a saved log lets you compare two incidents side by side. I do this on every support call now.
One more thing worth stating plainly: the System log tells you what happened, not why in a human sense. You still have to connect the dots between process names, scheduled tasks, and hardware. Event Viewer is a good clue sheet, not a verdict. But it turns a random, scary reboot into a specific named suspect, and for an office machine that is usually enough to fix it.