If Cloudflare goes down, half the internet goes down

If ChatGPT or Twitter (X) isn’t working for you, you’re not alone. In the last few hours, there have been problems at Cloudflare that have ended up affecting a multitude of platforms that use its services. This company, as happens with Amazon Web Serviceshas become one of the pillars of the internet, and when Cloudflare has problems, half the internet is infected.

Another fall. It’s not the first and it probably won’t be the last time we experience a Cloudflare outage. This company has a gigantic content distribution network (CDN, for Content Delivery Network) that is used by a multitude of internet services and platforms. This infrastructure allows websites of all types to be easily and quickly available to users, but a failure in its operation causes crashes like today’s.

When accessing services like X.com (Twitter) we now encounter an error message. The cause is clear: Cloudflare.

What they say on Cloudflare. Those responsible for the company have a status website for your infrastructure. In it we can see how the company currently indicates that they are aware and are “investigating a problem that affects multiple clients.” Just before that problem notice they were showing several maintenance operations messages on some of their nodes (Atlanta, Los Angeles, Tahiti) today. At 12:17 (Madrid time) they already warned of a problem that, however, may not have to do with this last one at 12:48.

Twitter (X) and ChatGPT, among those affected. The fall of Cloudflare has left platforms such as Twitter (X) or ChatGPT without service, which either do not work or do so irregularly. It was even difficult to check which other services were down because platforms like DownDetector, which precisely allows you to detect affected websites, was also affected by the Cloudflare downfall.

It’s time to be patient. As often happens in these cases, users can do little beyond being patient. Cloudflare is still investigating the problem at this time, but both Twitter and Downdetector and other affected services seem to be returning to normal.

Update (13.38): Cloudflare indicates that services are beginning to recover but “higher than normal” error rates continue to be observed. They continue to investigate the problem.

In Xataka | How a mistake in the network configuration of a small US business brought down Cloudflare, Facebook, Amazon and others globally