Cloudflare said that its Sunday morning shutdown involving various websites was due to an IP interruption by the Internet service provider CenturyLink. According to a tweet from CenturyLink, all affected services have been restored to 11:15 AM ET.
Today we have seen a massive online internet outage that has affected several providers, the Cloudflare representative said in an email to The Verge. There was not a particular Cloudflare outage. Level 3 / CenturyLink was blamed for an outage involving many Internet providers, including Cloudflare. The automated systems of Cloudflare recognized the problem and routed around it, but the magnitude of the problem still needed manual intervention.
The shutdown started early Sunday, and according to Cloudflare's status report, an elevated number of HTTP 5xx class errors, such as 522 & 503, were observed. In a later update, Cloudflare said that it had found a "transit provider problem" as the cause of the issue, and that it was working on a patch. Both data centers that use the transit service were impacted by the outage.
Cloudflare 's systems are planned to discourage websites from suffering interruptions due to high traffic loads as well as DDoS or spam comments attacks. DownDetector revealed that Sunday's blackout was affecting the US and parts of western Europe, and that sites like Discord, Hulu, Feedly, Xbox Live, and hundreds of others seemed to be affected.
CenturyLink experienced a major disruption in 2018 involving 911 calls, ATM deposits, lottery draws and other facilities.