One of the biggest cultural changes I had when I jumped the fence from Windows to Linux was the reboot. On Windows, the culture was strongly "reboot first, if it still happens, then it's a problem". On linux, it was a measure of last resort, because then you can't fix the problem (as you say).
Obviously the latter is the best way, but it's interesting that the culture of the two systems is so different, no doubt borne from Windows' legacy era.
There are windows sysadmins who schedule nightly server reboots "to keep the machines healthy". I've seen them in action. It is scary. The scariest part? It helps.
Except you have to triage. I'll happily reboot any box which fails unusually because my overall system should be HA enough to survive that. It's only interesting if the failure count is high, and I'm running reasonably up to date code. Otherwise there's too many things in a day to get through.
I'm with Vacri; I'm very reluctant to do that. Small problems can be harbingers of big ones. Even when they aren't, small problems often confound the ability to solve big ones.
Most concerning for me, though, is what safety experts call "normalization of deviance". It's the process by which people become accustomed to small failures, which creates opportunities for big failures to happen. A big example is the Challenger disaster. [1]
I see shops with low bug rates, where people think a lot about quality. And I see shops that, thanks to high bug rates, are too busy fighting fires to ever spend much time on quality. I never see any place in between. And I think normalization of deviance is why.
Obviously the latter is the best way, but it's interesting that the culture of the two systems is so different, no doubt borne from Windows' legacy era.