Technology and Science05 December, 20252 min read
Cloudflare Knocks Zoom, LinkedIn, Twitter and TfL Offline in Global Outage
This article is an automated aggregation of coverage from news outlets. Our editorials offer in-depth analysis curated by the NewsCord team.
Cloudflare outage knocked LinkedIn, Zoom, Twitter and Transport for London offline worldwide Cloudflare investigated Dashboard and API failures, implemented a fix, and restored services
7 outlets told this the same way.
Cloudflare outage overview
A major Cloudflare outage on Friday morning knocked large swathes of the internet offline for roughly 20 minutes (about 08:52–09:13 GMT/UK). Users were confronted with empty pages and service errors across many high-profile sites.
“A major Cloudflare outage this morning took many high-profile sites offline for about 20 minutes (approximately 08:52–09:13 GMT), affecting services including X/Twitter, Substack, LinkedIn and Canva.”
Reported disruptions included X (Twitter), Zoom, LinkedIn, Canva, Substack, Shopify, Coinbase, Discord, Deliveroo, TfL and Vinted, and monitoring site DownDetector briefly showed errors while logging thousands of reports. Many affected services were reported back online after the short disruption.
Cloudflare outage explanations
Cloudflare and its technical leads told reporters the incident was not being treated as an attack and pointed to recent configuration or maintenance changes as the proximate cause. The Mirror quoted Cloudflare's CTO Dane Knecht saying the issue was not being treated as an attack and linked the failure to disabling some logging while mitigating a recent React CVE. The London Evening Standard said engineers traced errors to Workers and Workers KV, citing empty pages from the list API and increased Workers script errors after that logging was disabled.
The Associated Press reported an alternative but related explanation: a firewall request-handling change and a database change during planned maintenance 'went slightly awry' and overloaded systems. The Shropshire Star said Cloudflare reported problems with its Dashboard and related APIs.
Outage impact summary
The outage's immediate visible effects varied by observer: users encountered blank pages and hosting or connection errors. DownDetector registered anywhere from a few thousand to more than 4,500 reports, depending on the source.
“Downdetector logged over 2,000 reports but undercounted the problem (itself briefly showed a 500 error).”
At least one regional company, India-based broker Groww, reported transient problems tied to the global disruption. DownDetector itself briefly returned a 500 error even as it tracked incident reports.
Cloud outage implications
Reporting placed the outage in a broader pattern of cloud-provider incidents and urged steps to reduce single points of failure. The Evening Standard noted that Cloudflare serves an estimated 20% of websites and urged firms to consider multiple cloud, CDN and authentication providers.
The Mirror and Associated Press framed the disruption as part of a series of recent high-profile outages and quoted lawyers and experts warning about systemic risk and potential regulatory scrutiny under rules such as DORA, NIS2 and new UK operational resilience requirements. The Shropshire Star reiterated that many affected sites have since come back online.
Cloudflare outage response
Cloudflare executives promised a full post‑mortem and fixes to reduce recurrence. The Evening Standard reported the CTO and CEO vowed a full post‑mortem and changes to prevent a repeat, with Matthew Prince describing planned fixes. The Shropshire Star said Cloudflare had implemented a potential fix and was monitoring the results.
“The company’s CTO and CEO promised a full post-mortem and changes to prevent recurrence; Matthew Prince described a detailed recount of failures and planned fixes.”
AP quoted an outside expert who said a database change "went slightly awry," highlighting the vulnerability of centralised infrastructure. The companies and outlets agreed on the immediate technical fault and on commitments to explain what went wrong and prevent a recurrence.
