Cloudflare has apologised for an outage on Friday morning that disrupted websites including LinkedIn, Zoom and Downdetector, marking its second service failure in less than a month. The company stated that any outage is unacceptable and acknowledged letting the internet down again.
The outage, which affected 28% of Cloudflare's traffic, lasted for half an hour and was resolved shortly after 9am GMT. Cloudflare attributed the issue to a firewall adjustment made to protect customers from a widespread software vulnerability, and confirmed it was not an attack. A separate problem with its application programming interfaces was also reported.
This incident follows a larger Cloudflare outage in mid-November that hit platforms such as X, OpenAI and Spotify. That outage was caused by a configuration file growing beyond its expected size, triggering a crash. Cloudflare has since spoken with hundreds of customers and shared plans to prevent similar widespread impacts.
Experts have raised concerns about the centralisation of internet services. Professor Steven Murdoch of University College London noted that companies may start questioning Cloudflare's reliability after two outages in quick succession. Michał Woźniak, a DNS expert, described the recent outages as evidence of how brittle big tech infrastructure has become, calling it the fourth major global outage since October.
Cloudflare serves about 20% of all websites and has nearly 300,000 customers across 125 countries. Despite the outages, some experts suggest the incidents could inadvertently highlight Cloudflare's widespread use, potentially serving as a form of marketing.
