Even occasional Twitter users will no doubt be familiar with the service’s frequent downtime. It’s a rare day when I don’t run into at least one or two “something’s technically wrong” messages on the site. That has prompted a lot of discussion about how to improve Twitter, and also some discussion about how things could be so bad.
I’ve been willing to cut them some slack. They’ve grown exponentially, and continue to do so. Then on Wednesday, Twitter founder Jack posted this on the official blog:
We’ve gone through our various databases, caches, web servers, daemons, and despite some increased traffic activity across the board, all systems are running nominally. The truth is we’re not sure what’s happening. It seems to be occurring in-between these parts.
Transparency is great, but surely they must have some idea about what’s wrong? I don’t know much about their architecture or systems, but it seems odd to me that they’d be totally stumped. It suggests to me that their architecture was never designed, and was instead thrown together over time. Now they’re in too deep to start over.
Twitter developer Alex suggests that the main problem is the system was originally put together as a content management system, when in reality it’s a messaging system. If that’s the case, fine, but messaging systems are not new. They must be able to examine and learn from some existing stuff right?
Posts like the one Jack made don’t inspire much confidence that they’ll be able to turn things around, but I sure hope they do. I really love Twitter. Maybe the $15 million in additional funding that they recently secured will help.