On November 18, 2025, Cloudflare, one of the world’s largest content delivery networks (CDN), DDoS protection providers, and DNS services, experienced a significant internal service degradation that triggered widespread internet disruptions. The incident began around 11:48 UTC (approximately 6:48 AM ET / 3:48 AM PT) and affected thousands of websites, apps, and online services globally. Users encountered frequent “500 Internal Server Error” messages, CAPTCHA challenges failing to load, or complete inaccessibility on platforms reliant on Cloudflare’s infrastructure.
Cloudflare powers security, performance, and reliability for an estimated 20% of the web’s websites (based on industry reports), making outages of this scale highly impactful but not a complete “global internet blackout.” While no official percentage of the entire global internet was reported as offline—Cloudflare does not route all internet traffic—the disruption rippled through a substantial portion of popular online services, highlighting the interconnected fragility of modern web infrastructure.
Timeline of the Outage (All Times in UTC)
- 11:48 UTC: Cloudflare first acknowledges issues with its support portal provider, noting errors in viewing or responding to support cases. Simultaneously, reports of widespread 500 errors emerge, affecting the Cloudflare Dashboard, API, and customer traffic.
- 11:20–12:00 UTC: A spokesperson later confirms a “spike in unusual traffic” to one of Cloudflare’s services began around 11:20 UTC, triggering errors across the network.
- 12:03 UTC: Official status update: “Cloudflare is experiencing an internal service degradation. Some services may be intermittently impacted.”
- 13:09 UTC: Issue identified; a fix is being implemented.
- 13:13 UTC: Changes applied allow Cloudflare Access and WARP (zero-trust security tools) to recover. Error rates return to normal for these services; WARP access re-enabled in London.
- 13:35 UTC: Ongoing work to restore other application services. Higher-than-normal error rates persist for some customers.
- Later Updates (as of ~14:00 UTC): Services show signs of recovery, but intermittent issues continue. No full resolution announced yet, though outage reports on trackers like Downdetector begin declining.
The outage lasted several hours at peak intensity, with partial recoveries starting around 13:00 UTC. Cloudflare noted scheduled maintenance in datacenters (e.g., Santiago, Chile from 12:00–15:00 UTC), but it remains unclear if this contributed.
Cause of the Outage
Cloudflare has described the incident as an “internal service degradation” stemming from a “spike in unusual traffic” to one of its services starting at approximately 11:20 UTC. This led to cascading errors in traffic processing across its global network.
- No evidence of a cyberattack (e.g., DDoS) has been reported; it appears to be an internal issue rather than external malice.
- Some speculation linked it to ongoing scheduled maintenance in locations like Santiago (SCL), Atlanta, Los Angeles, and Tahiti, but Cloudflare has not confirmed this as the root cause.
- Unlike past outages (e.g., a June 2025 incident blamed on a third-party dependency or backbone failures), this one involved widespread 500 errors and challenges (e.g., CAPTCHA failures), affecting bot protection and traffic routing.
A full post-mortem report from Cloudflare is expected in the coming days, as is standard for major incidents.
Scope and Impact: How Much of the Internet Was Affected?
No precise “percentage of global internet outage” was provided by Cloudflare or independent monitors like Cloudflare Radar or Kentik. However:
- Cloudflare handles traffic for ~20% of websites worldwide (per historical industry estimates from W3Techs and similar sources).
- The outage was not global in the sense of taking down the entire internet—core internet routing (e.g., via Tier 1 providers) remained functional. Instead, it caused intermittent or complete failures for sites using Cloudflare’s CDN, DNS, security (e.g., WAF), or Zero Trust services.
- Peak reports on Downdetector exceeded 5,000–11,000 for Cloudflare itself, with tens of thousands more for downstream services. Even Downdetector briefly went offline due to its own reliance on Cloudflare.
- Affected regions: Issues reported globally, with notable impacts in Europe (e.g., nodes in Bucharest, Zurich, Warsaw down) and North America.
This was comparable to recent outages like AWS in October 2025, but more targeted to web-facing services rather than cloud computing backends.
Major Sites and Services Affected
The outage hit a diverse range of platforms, as many unrelated sites share Cloudflare as a common dependency. Confirmed or heavily reported impacted services include:
| Category | Affected Services/Platforms | Symptoms Reported |
|---|---|---|
| Social Media | X (formerly Twitter), Facebook/Meta services | Inaccessibility, loading failures |
| AI/Chatbots | ChatGPT (OpenAI), Claude AI | Login failures, 500 errors |
| Gaming | League of Legends, other multiplayer titles | Server connection issues |
| Productivity | Canva, Discord, Spotify | Partial or full downtime |
| Dating/Shopping | Vinted, bet365 | Blocked access, errors |
| Other | Letterboxd, Downdetector (ironically), various banking apps, crypto platforms, NJ Transit app | 500 Internal Server Errors, CAPTCHA challenges failing |
Thousands of smaller sites, enterprise dashboards, and APIs also failed, affecting businesses’ ability to manage their Cloudflare configurations during the incident.
Broader Implications
This event underscores the risks of centralized infrastructure: a single provider’s issue can cascade across competitors (e.g., X and OpenAI both down simultaneously). It follows a pattern of 2025 outages, including AWS disruptions and earlier Cloudflare incidents.
Cloudflare’s engineering teams responded quickly, prioritizing critical services like Access and WARP. As of the latest updates, most traffic is recovering, though users may still see elevated errors.
For real-time status, check cloudflarestatus.com. Cloudflare typically publishes detailed incident reports on their blog post-outage.
This outage serves as a reminder for websites to consider multi-CDN setups or failover mechanisms to mitigate single points of failure in the future.
LogicWeb and global customers remain online despite Cloudflare outage.
