So Wikipedia went down today on account of a tripped circuit breaker behind the UPS of the colocation facility. WTF? This is even weirder than when more-or-less-exactly the same thing happened to LiveJournal last month. LJ's problem was that someone inside the server room hit an emergency shutdown switch when they shouldn't have - but the hardware did exactly what it was supposed to, someone was just a dumbass. At Wikipedia, either the hardware was installed with monumental stupidity or it was broken in a way that can only be described as a hideously un-robust single point of failure.

I have no idea how either of these events managed to occurr. That's the whole point of these giant facilities - they worry about this stuff for a living. Wow!