Part of the internet was unavailable on Tuesday afternoon due to a Cloudflare outage. Accessed sites returned a 502 Bad Gateway message. CEO Matthew Prince confirmed on Twitter that Cloudflare had a network problem. Now the problems seem to be over.
The outage started in mid-Tuesday afternoon, Cloudflare reported having “Network Performance Issues” on a System Status page. Discord, among others, was therefore not reachable. At 3:52 p.m. Cloudflare wrote to investigate the issue and at 4:15 p.m. the company reported implementing a solution and seeing if it worked. The company does not provide any further information.
The solution seems to have worked. For example, raspberrypi.org was unavailable on Tuesday afternoon, but at the time of writing the site is back online. Cloudflare CEO Matthew Prince confirms on Twitter to have ‘softened’ the problems and that network traffic is working again. In an update he reports that CPU spikes were the cause of the failure of systems and backup systems and the resulting impact on all services. He said there are no indications that it was an attack.
The influence of the outage on the internet was even visible in the statistics of internet node AMS-IX: there was a significant dip in the traffic that the node was processing.