November 18, 2025 — A large-scale outage at Cloudflare caused widespread disruptions across the internet on Tuesday, affecting numerous major websites and applications that rely on the company’s infrastructure.
The outage began late morning UTC, when Cloudflare systems experienced a spike in abnormal traffic that led to elevated error rates across its network. As a result, millions of users around the world encountered connection failures and 500-level server error messages on platforms that depend on Cloudflare for security, content delivery, and DNS services.
Services affected included major platforms such as ChatGPT, X (formerly Twitter), and several online gaming and streaming services. Many smaller websites and online businesses that rely on Cloudflare’s network also reported complete outages, leading to downtime across multiple sectors.
Cloudflare confirmed it was experiencing what it described as “internal service degradation.” Engineers worked throughout the day to stabilize systems and implement fixes to restore normal functionality. Some services, including Cloudflare Access and the WARP connectivity service, began returning to normal levels in certain regions as recovery efforts progressed.
Despite partial restoration, many customers continued to encounter intermittent issues as Cloudflare investigated the root cause of the disruption. Early indications suggest the incident was triggered by unexpected traffic surges rather than an external attack, though the company has not released full technical details.
The outage highlights the critical role Cloudflare plays in delivering internet traffic globally. With millions of websites and applications dependent on its infrastructure, disruptions of this scale can have cascading effects across the digital ecosystem. Businesses relying on Cloudflare experienced service downtime, and users across multiple continents reported widespread accessibility problems.
Cloudflare stated that engineers will continue working to fully resolve the incident and restore system stability. A post-incident analysis is expected to be published once the investigation is complete.