What’s the skinny? A rather large network outage lasting several hours occurred overnight last night. In an email sent out to it’s customers, Cloudflare stated that: “at approximately 10:03 UTC, a misconfiguration at a major external network provider caused connectivity disruptions across the Internet. These disruptions impacted the operations of many service providers, including Cloudflare.”
Thanks to their automated system only a few minutes went by before the system automatically rerouted traffic away from CenturyLink. Still however some customers continued to see issues which Cloudflare attributes to CenturyLink being so large. “CenturyLink/Level(3) is among the largest network providers in the world. As a result, many hosting providers only have single-homed connectivity to the Internet through their network.”
The graph below displays the six major tier-1 networks Cloudlfare connects to and shows CenturyLink in red dropping to near-zero during the incident.
The outage lasted approximately four and a half hours and Cloudflare says they’re still awaiting a postmortem from CenturyLink. The email I received from Cloudflare however suggested that a misconfiguration on CenturyLinks end is to blame. I suppose that’s just a blanket statement they’re using as they await news of the real “cause of death”.