[nothacker]

Redundancy wasn't the problem I saw last night. What I saw, at least with Heroku, is that when I checked, the main Heroku site was down and displaying things like nginx errors. That to me is unacceptable for an operation such as theirs. Even if all hell is breaking loose, you don't only keep your status page up for all to see, you have a pretty damn good message up that the main page resolves to. I'm not saying they screwed the pooch entirely as I'm sure they were busy, but, damn it, even Amazon is going to go down sometimes. Screw redundancy if you can't even serve a webpage to inspire confidence that you are working on it. I'm sorry I'm picking on Heroku specifically, because I'd be really f'n surprised if a lot of you weren't in the same boat. You need to have the main page served when that happens, even just a static page that inspires confidence or direct to the blog and provide updates there.