A "cascading" infrastructure glitch forced popular social network Twitter offline last night, for several hours from around 5pm GMT.
Twitter struggled to restore service, announcing success about an hour after the outage began, only to have the network go down again shortly after that.
Explaining the down time in a blog post, Twitter engineer Mazen Rawashdeh said that the outage was down to "a cascading bug in one of our infrastructure components".
"This wasn’t due to a hack or our new office or Euro 2012 or GIF avatars, as some have speculated today," Rawashdeh wrote. "A ‘cascading bug’ is a bug with an effect that isn’t confined to a particular software element, but rather its effect “cascades” into other elements as well.
"One of the characteristics of such a bug is that it can have a significant impact on all users, worldwide, which was the case today."
To get rid of the bug, Twitter was rolled back to a previous, stable version.
Twitter’s underlying IT architecture is constantly evolving. In a blog post at the end of May, the engineering team explained some of the changes to the service.
When the new version of Twitter was released in September 2010, the web client worked by receiving an API from Twitter’s servers, then rendering a user interface on top of that using javascript, right in the browser. To improve performance, Twitter started to do the rendering on the server side, using javascript in the browser just to provide interactivity.
At the end of the post, engineering manager said Dan Webb said that Twitter would soon start overhauling the server side of the application.