Major Cloudflare global outage leaves ChatGPT, X and dozens of sites intermittently offline after “unusual traffic spike”
On November 18, 2025, a global Cloudflare outage caused by an “unusual traffic spike” triggered widespread HTTP 500 errors, intermittently taking services such as ChatGPT and X offline while teams worked to mitigate and restore global traffic.
- Spike triggered HTTP 500 errors: Cloudflare reported an “unusual traffic spike” around 11:20 UTC that overwhelmed parts of its edge network and control plane.
- Broad impact: High-profile services including ChatGPT, X, Canva and Downdetector experienced intermittent outages; Cloudflare’s dashboard and API were also affected.
- Mitigation and partial recovery: Initial mitigations were implemented by 13:09 UTC with wider restorations beginning from 13:13 UTC as fixes propagated.
- Investigation pending: Cloudflare said the root cause was unknown and that teams were “all hands on deck” to restore service and then investigate.
What happened
Cloudflare reported a major network incident after an unusual spike in traffic began at about 11:20 UTC, causing HTTP 500 errors and failures across parts of its global network. The incident affected many customer sites and some Cloudflare services, including the dashboard and API. Reporting and updates were provided by Tom’s Hardware and Cloudflare status.
Which services went dark, and how users saw it
Users worldwide reported problems loading websites protected by Cloudflare. ChatGPT users saw failures reaching the AI chat service, while X users experienced errors posting and viewing timelines. Other affected platforms included Canva, Windscribe VPN and even outage trackers such as Downdetector, which limited real-time visibility of the disruption. Independent coverage includes The Independent and ABC News.
What Cloudflare reported
Cloudflare’s status updates described a global outage driven by an “unusual traffic spike”. The company said teams were deploying mitigations across data centers and acknowledged partial impact to its control plane — the dashboard and API — complicating customer workarounds. Cloudflare stated teams were focused on serving traffic without errors and would investigate the root cause once service stabilized (Cloudflare status).
“We do not yet know the cause of the spike in unusual traffic. We are all hands on deck to make sure all traffic is served without errors. After that, we will turn our attention to investigating the cause of the unusual spike in traffic.” — Cloudflare status updates (as reported by Tom’s Hardware)
Timeline and remediation (detailed)
- 11:20 UTC: Cloudflare observed a rapid and unusual increase in traffic that began causing HTTP 500 server errors for many customers (Tom’s Hardware).
- 11:20–13:00 UTC: Errors continued and spread across regions; Cloudflare posted updates while working to identify and mitigate the issue (Cloudflare status).
- 13:09 UTC: Cloudflare implemented initial mitigations to reduce error rates and normalize traffic handling (Tom’s Hardware).
- 13:13 UTC onward: Some services, including Cloudflare Access and WARP, began to recover for many users as fixes propagated; residual error rates remained during global propagation (Cloudflare status).
Why this mattered: the internet is tightly linked
Cloudflare is a major internet infrastructure provider whose network provides content delivery, security and performance features for many websites. When Cloudflare’s systems have problems, the effects can cascade quickly across services that rely on it. The outage highlights how a small number of central providers can create single points of failure even as they offer global scale and performance benefits.
Customer and vendor responses
OpenAI pointed to problems at a third‑party provider as the reason for its ChatGPT interruptions. Many affected companies acknowledged service interruptions via their channels, though communication was sometimes incomplete because Cloudflare’s dashboard and status systems were partly affected. Independent reporting and social media served as key real‑time information sources for users (Tom’s Hardware; The Independent).
Context on maintenance and next steps
Cloudflare had performed scheduled maintenance earlier the same day. Company updates said no confirmed link had been established between that maintenance and the traffic spike; a post‑incident investigation and detailed root‑cause analysis were promised once immediate impacts were resolved (Tom’s Hardware).
Implications: economic, political and social
Economic impact: Short outages at major infrastructure providers can cost businesses in lost sales and productivity — affecting e-commerce, media, education and finance. Small businesses using SaaS tools protected by Cloudflare may have faced interrupted orders and bookings.
Political consequences: Policymakers increasingly focus on digital resilience and critical infrastructure. Widespread outages affecting public debate platforms and news services can complicate election communication and emergency alerts, prompting calls for clearer reporting standards and more redundancy.
Social and cultural effects: Users reported frustration as familiar services became unavailable, disrupting work and study and renewing conversations about digital dependence, local backups and offline alternatives.
Practical applications for citizens, organisations and regulators
- For individuals and companies: Keep alternative contact channels, maintain simple offline procedures for critical tasks, and consider multi‑provider redundancy for essential services.
- For organisations: Review vendor dependency, evaluate multi‑CDN or multi‑provider strategies, and test failover procedures and communications plans.
- For governments and regulators: The incident reinforces calls for clear incident reporting, post‑incident analysis and transparency about testing and deployment practices for large infrastructure providers.
Reporting sources
- Tom’s Hardware: “Cloudflare outage under investigation as Twitter, Downdetector go down; company confirms global network issue”
- The Independent: Coverage of Cloudflare downtime and affected services
- ABC News: “Outage at Cloudflare disrupts access to popular websites”
- Cloudflare status: Real-time incident reporting from Cloudflare
Reporting by Times Media Service
