Cloudflare outage brings down major websites Including Twitter, ChatGPT, and Spotify
Learn More
Cloudflare, one of the world's largest internet infrastructure and content delivery network providers, experienced a major global outage on November 18, 2025, that disrupted access to hundreds of thousands of websites and online services for approximately six hours.
The incident began at 11:20 UTC when the company's global network started experiencing widespread service failures. Users attempting to access affected websites encountered "internal server error" messages displaying HTTP 500 status codes, rendering countless online platforms inaccessible during the disruption. Cloudflare acknowledged the issue at 11:48 UTC, warning customers that multiple services including its dashboard and API were failing, and stated it was investigating the problem that impacted numerous customers across its infrastructure.
The root cause of the outage was identified as a configuration file used by Cloudflare's Bot Management system that automatically generates and deploys machine learning model features to detect and manage bot traffic across the network. According to Cloudflare's official incident report, this feature configuration file, which is refreshed and published every few minutes to allow rapid response to evolving bot attacks and traffic patterns, unexpectedly grew beyond its expected size of entries. When the oversized configuration file was deployed across Cloudflare's global network, it triggered a crash in the software system responsible for routing and handling traffic for Cloudflare services.
The company claims there was no evidence that the outage resulted from a cyberattack, distributed denial-of-service attack, or any malicious activity, characterizing it instead as an internal technical failure in their automated threat management infrastructure.
The outage had massive cascading effects across the internet, impacting some of the world's most heavily trafficked websites and critical online services:
- Social media platforms: Twitter users were unable to post or access content
- Artificial intelligence services: OpenAI's ChatGPT and Sora video generation platform, Anthropic's Claude AI chatbot
- Streaming and entertainment: Spotify music streaming service, Letterboxd film review platform, League of Legends and Valorant gaming platforms
- E-commerce and business services: Shopify e-commerce platform, Canva design platform, Indeed job search engine, IKEA retail websites
- Financial and workplace systems: Dayforce workforce management platform, Truth Social
- Infrastructure monitoring: DownDetector itself, ironically preventing users from checking outage status
- Critical services: McDonald's self-service ordering kiosks, nuclear facility visitor access systems (PADS), public transit digital services
utage monitoring service DownDetector received tens of thousands of user reports within minutes of the incident beginning, with peak reports exceeding 11,000 simultaneous complaints.
Cloudflare's engineering teams worked throughout the morning to identify and remediate the issue. The company provided multiple status updates as recovery progressed.
- At 12:21 UTC, Cloudflare reported seeing initial signs of recovery with services beginning to come back online, but customers continued experiencing higher-than-normal error rates.
- By 13:13 UTC, the company confirmed that changes had been deployed to restore Cloudflare Access and WARP services. Error rates returned to pre-incident levels. Application services remained impacted during this phase as engineers continued working on restoration.
- The incident was officially declared resolved at 17:44 UTC, approximately six hours after it began. Cloudflare confirmed that services were operating normally and no longer experiencing elevated errors or latency across the network. The company says it would perform deeper investigations into the failure but refrained to make additional configuration changes during the immediate post-incident period.
Cloudflare Chief Technology Officer Dane Knecht issued a public apology acknowledging the severity of the failure and its impact on customers and the broader internet ecosystem.
Cloudflare accepts that it let customers down and committed to conducting comprehensive post-mortem analysis to understand all system and process failures that contributed to the incident.
