Someone, somewhere probably just got fired for this massive cock-up… CloudFlare – a content delivery network that effectively speeds up websites (including this one) and decreases the effects of geographic slow-down failed this morning, bringing down a large chunk of the internet with it.
At exactly 09:47:31 AM (GMT) I was woken (yes I know that’s late, but it is Sunday) by an alarm system I have in place that alerts me if JHACK goes offline. Quickly checking to see what the problem was, I noticed that it was a DNS issue and that CloudFlare would probably be to blame. This has happened before, and normally only lasts a few minutes at most. To my surprise though, the outage lasted a whopping 1 hour and 4 minutes! What’s worse is that this wasn’t a localised incident, the entire global network was down, meaning that 4Chan, imgur and over 800,000 other websites were all unreachable.
We’re still waiting on a blog update from CloudFlare but at the moment it’s most likely that the issue had something to do with scheduled maintenance that happened yesterday. It failed meaning that around 75% of CloudFlare sites were momentarily down. The likely cause for today’s serious outage is that something broke, probably a routing mechanism, and brought down all CloudFlare sites. My money’s on this being related to yesterday’s issues, time will tell through.