WP.com downtime summary
en.blog.wordpress.com
en.blog.wordpress.com
If you're gonna have an unscheduled outage, at least apologize the right way.
(P.S. 5.5m PVs in 110 minutes is a little over 800 req/sec. That's much lighter than I would have expected from the entire Wordpress.com empire, but I guess it was the middle of the day rather than prime-time)
On the other hand, if equipment is routinely powered down and or unplugged, the technicians working on it will have a better idea of what goes where, which gizmo affects what doodad, etc. It just generally seems like a good idea to do this under controlled conditions rather than unexpected and uncontrollable ones.
I personally would rather have a 95% uptime with close to 5% scheduled downtime than 99% uptime with 1% unscheduled downtime. The former scenario lets me set clear expectations to customers. The latter doesn't. I can't tell them what the hell is going on or when it will get better. It often takes days to find out why it even ended up happening if an explanation is ever discovered.