How Facebook (briefly) killed the Internet

For a few minutes Thursday evening, Facebook was redirecting users visiting dozens of websites – including Mashable and CNN – to cryptic error pages.

The reaction online was pretty much what you’d expect, with – as The Next Web noted – hashtags like “Facebookmageddon” and “Facebocalypse” common amongst Twitter users.

So what happened, exactly? There was an issue with the Facebook Connect API that caused users on sites that use that API to redirect users to a Facebook error page.

For example, if you were visiting Mashable and logged into our site using your Facebook account (and you were also signed into Facebook), you were automatically redirected to a Facebook error page.

Exiting the page or attempting to re-access the original site would lead to another.

Sites such as The Huffington Post, Kayak, Hulu, The Daily Dot, Pinterest and hundreds of others were all impacted. The bug lasted less than 10 minutes.

In a statement, Facebook told Mashable: “For a short period of time, there was a bug that redirected people logging in with Facebook from third party sites. The issue was quickly resolved and Login with Facebook is now working as usual.”

The bug may have been brief, but it has highlighted just how many important websites use Facebook Connect for user authentication.

Over the span of just a few years, Facebook logins have become so pervasive that they are nearly second nature. It also shows that if Facebook has an issue, it can affect more than just its site – it can also impact the hundreds of thousands (millions?) of sites that integrate with Facebook’s APIs.

What’s interesting is that a user didn’t even need to be performing the action for the error – and hijacking – to occur. Instead, simply being logged into both places (and having the accounts linked) was enough to force users off of a third-party website and onto Facebook’s error page.