On November 18, 2025, Cloudflare, a major web infrastructure provider handling about 20% of global internet traffic, suffered a widespread outage starting around 11:20 UTC. The incident caused 500 internal server errors across numerous sites, affecting millions of users worldwide.
Cloudflare reported an “internal service degradation” and a spike in unusual traffic to one of its services. This led to failures in its dashboard, API, and core network functions. Ironically, even Cloudflare’s status page and Downdetector were impacted initially.
Major platforms hit included X (formerly Twitter), OpenAI’s ChatGPT, Spotify, Facebook, Canva, Perplexity, Letterboxd, League of Legends, and crypto sites like DefiLlama. Users saw error messages blaming Cloudflare, with spikes in outage reports on tracking tools.
The timing coincided with scheduled maintenance in data centers like Santiago (Chile), Los Angeles, Atlanta, and others, though Cloudflare hasn’t confirmed a direct link. Engineers quickly identified the issue, deployed fixes, and reported recovery for services like Cloudflare Access and WARP by early afternoon UTC. Full restoration was ongoing, with lingering elevated error rates.
This event echoes recent disruptions, such as AWS’s October 2025 outage and Microsoft’s Azure issues, highlighting the internet’s vulnerability to single points of failure in centralized providers.
Cloudflare has a history of outages: a global one in March 2025 from credential errors, R2 storage failure in February 2025, and earlier incidents in 2024 from power failures or config changes. Each prompts post-mortems to improve resilience.
The outage underscores over-reliance on few infrastructure giants, urging diversification for critical services. Cloudflare shares dropped over 5% in premarket trading amid the chaos.
As services stabilized, it served as a reminder of the fragile backbone powering the modern web.