Updated Routine Cloudflare maintenance went awry this morning, knocking over the company’s dashboard and API and sending sites around the world into error screens.
Cloudflare was working through its scheduled servicing when things went sideways. Maintenance was in progress in its Chicago datacenter from 0700 UTC, with work due to begin in its Detroit datacenter at 0900 UTC when red lights began flashing at administrators around the world.
Cloudflare status this morning
The content delivery network giant admitted a problem with its service at 0856 UTC, rolled out a fix shortly after, and seemed to be back up and running by 0930 UTC. It has, however, now reported issues with Workers (the serverless functions, not the employees likely frantically trying to stop the company’s systems from falling over again).
Cloudflare on Down Detector
We’ve asked the company for more information, and will update this piece should an explanation be forthcoming.
Cloudflare proudly proclaims that “20 percent of all websites are protected by Cloudflare.” Unfortunately, this also means that 20 percent of all websites could catch a cold should Cloudflare sneeze. Two outages in two months is less than ideal, and could cause affected customers to take a hard look at their dependencies.
High-end UK retailer Fortnum & Mason, for example, was one such casualty. It showed only a sorrowful “500 Internal Server Error” instead of the goodies that Christmas shoppers expect to find this time of year. The site has now recovered, but the simple text “cloudflare” that was shown beneath the error is not a good look for the internet services company.
Web page showing a server error during Cloudflare outage
Social media was its usual sympathetic self as users wailed that their favorite website had stopped responding.
The outage does not appear to be as long-lasting as the one in November, which took out a substantial chunk of the web, thanks to an ill-conceived database permissions change. ®
Updated at 12.01 UTC on December 5, 2025, to add:
A spokesperson at Cloudflare sent us a statement after publication:
“A change made to how Cloudflare’s Web Application Firewall parses requests impacted the availability of Cloudflare’s network at approximately 8:47 GMT and concluded approximately 9:13 GMT. This was not an attack; the change was deployed by our team to help mitigate the industry-wide vulnerability disclosed this week in React Server Components.”
