A little unplanned maintenance to take care of this evening. We should be back online in just a few minutes.
Wow, you guys blip a LOT. We tracked the problem back to a corrupt table in the master database that stores all the blips that people post. Don’t worry, they’re safe, but due to the total number of blips it takes a while to get things back in order.
We’re already planning some changes that will keep the site up in case of failures like these, but as luck would have it we got hit with the ghost in the machine just a few days before we had planned to move over to the new setup.
If anyone is interested in the more nerdy side of running things, I plan to write a series of posts about the kind of hurdles and pitfalls you might run into when you’re building an internet company. There seems to be a common path from initial release through to scale that you don’t find out about until you’re dealing with it.