On Tuesday, November 18, Cloudflare experienced a major outage that disrupted access to numerous popular websites and services, including X and ChatGPT.
The issue began around 11:30 GMT and affected Cloudflare’s global traffic handling due to a bug in its Bot Management system. A configuration file doubled in size unexpectedly, causing a critical crash in Cloudflare’s core proxy software that manages internet traffic. This led to widespread HTTP 5xx server errors, resulting in site outages and login failures.
The company resolved the problem by reverting to a previous version of the configuration file and manually restarting affected services. Full recovery took several hours, with normal performance returning by 17:00 GMT.
During the outage, additional services such as Grindr, Zoom, Canva, and even Cloudflare’s own dashboard also experienced interruptions or slowdowns.
Cloudflare confirmed that the outage was caused by a technical fault and was not the result of any cyberattack. The company apologised for the disruption, acknowledging the unacceptable impact given its integral role in internet infrastructure. Share prices dropped by around 3% following the announcement.
Cybersecurity experts warn that outages at major services like Cloudflare, Amazon Web Services, and Microsoft Azure expose the fragility within today’s digital ecosystem, where few alternatives exist for hosting and security.
Cloudflare has committed to strengthening its systems to prevent similar failures in the future and restore confidence in its critical services.