Cloudflare Outage Sparks 500 Errors, Multiple Sites Offline

Why Cloudflare’s proxy is returning 500 Internal Server Errors and disrupting AI services like Claude and Perplexity —…

·MŽMatija Žiberna·
Cloudflare
Cloudflare Outage Sparks 500 Errors, Multiple Sites Offline

☁️ Cloudflare Edge Development Guides

Complete Cloudflare guides with practical examples, deployment strategies, and developer prompts to help you build and ship edge applications faster.

No spam. Unsubscribe anytime.

UPDATE (10:15 AM CET): As of this time, it appears the issue has been at least partially resolved. You can follow the current status here: https://www.cloudflarestatus.com/

Cloudflare is currently returning 500 Internal Server Errors and causing multiple well known services to fail. The outage is affecting several major platforms, including:

  • Claude (anthropic.com)
  • Perplexity AI (perplexity.ai)
  • Cloudflare's own website (cloudflare.com)
  • Downdetector (completely offline—the service meant to track outages)
  • KLAVIYO
  • Drip.com
  • BREVO
  • And many more

These platforms rely on Cloudflare to handle traffic and security at the edge. When Cloudflare misbehaves at that layer, the services above it cannot respond, even if their origin servers are fully functional.

This is the second large scale Cloudflare disruption in less than a month. The previous incident crippled DNS routing and took a significant portion of the web offline. This time the problem appears more targeted. Sites using Cloudflare only for DNS still load, suggesting the failure is happening in Cloudflare's internal request handling or proxying components rather than in its DNS backbone.

The symptoms are consistent: domains resolve correctly, but once traffic reaches Cloudflare's network, it fails to be processed. That is why users see 500 Internal Server Errors on Claude and blank pages or timeouts on Perplexity AI.

Outages like this can stem from several internal issues, such as:

  • A misconfigured rule or update pushed globally
  • A software bug in Cloudflare's routing or proxying stack
  • A critical internal service failing under unexpected load
  • A dependency inside Cloudflare's infrastructure malfunctioning

Because Cloudflare sits between users and application servers, a single point of failure inside its network can instantly take multiple unrelated services offline.

The frequency of outages raises concerns about Cloudflare's deployment processes and safeguards. With an increasing number of AI products depending on Cloudflare for performance and security, each incident now has wide visibility and broad impact.

This report will be updated once Cloudflare issues a statement or the affected services fully recover.

View markdown version
6

Comments

Leave a Comment

Your email will not be published

10-2000 characters

• Comments are automatically approved and will appear immediately

• Your name and email will be saved for future comments

• Be respectful and constructive in your feedback

• No spam, self-promotion, or off-topic content

Matija Žiberna
Matija Žiberna
Full-stack developer, co-founder

I'm Matija Žiberna, a self-taught full-stack developer and co-founder passionate about building products, writing clean code, and figuring out how to turn ideas into businesses. I write about web development with Next.js, lessons from entrepreneurship, and the journey of learning by doing. My goal is to provide value through code—whether it's through tools, content, or real-world software.