Cloudflare Crashed: The Biggest Internet Outage of 2025 Lasted 102 Minutes – It's Completely Over!

Cloudflare experienced a 102-minute outage on November 18, 2025! X, ChatGPT, Discord, Spotify, and Trendyol were inaccessible. The CEO apologized, and the BGP error was explained. Full details and timeline →

Cloudflare Crashed: The Biggest Internet Outage of 2025 Lasted 102 Minutes – It's Completely Over!

Cloudflare's global outage, which began today at 2:48 PM and lasted exactly 102 minutes, has officially ended. The company has just announced "Incident Resolved" and reported that all services are operational again.

This outage at Cloudflare, considered the "spinal cord" of the internet, rendered Twitter, ChatGPT, Discord, Spotify, League of Legends, Shopify, and tens of thousands of websites inaccessible simultaneously. While many users searched "Has the internet crashed?", what actually collapsed was one of the world's largest CDN and security infrastructures.

Outage Timeline

14:48 → The first packet loss and 500 errors began.
14:53 → Cloudflare Status page declared a “Critical Incident,” the map turned completely red.
15:05 → Official statement: “Critical issue – Dashboard and API are also not working”
15:18 → Incorrect configuration has been rolled back, fix has begun.
15:40 → Traffic has returned to 85%
16:30 → “All services operational” – Fully returned to normal
16:38 → The root cause has been explained.

The Root Cause: A Simple Yet Destructive Human Error

According to Cloudflare's initial post-mortem summary, the problem is as follows:

An incorrect prefix was declared in a BGP route (routing table) change made during a software update. This error created a "routing loop" on the global backbone → CPU usage suddenly jumped to 100% → all edge servers became unresponsive.

According to Cloudflare: “The misconfiguration that went into effect at 14:48 was withdrawn at 15:18. Full recovery was achieved at 16:30.”

This has gone down in history as an even longer and more disruptive incident than Cloudflare's 27-minute outage in 2019.

What happened in Türkiye?

Cloudflare traffic on Türk Telekom, Vodafone, and Turkcell backbones has dropped by over 90%.
Major e-commerce sites like Trendyol, Hepsiburada, and n11 reported 502/503 errors.
Garanti BBVA, ING, and QNB Finansbank experienced login problems with their online banking services.
Most news sites, forums, and Discord servers went offline.
As of now (17:10), all services in Türkiye have returned to normal.

Cloudflare CEO Matthew Prince Apologizes

Matthew Prince just tweeted the following:

“Today we made a huge mistake that affects millions of people and businesses. We apologize for this. The root cause was a simple human error. To prevent this from happening again, we are introducing much stricter control and automation mechanisms. A detailed technical report will be published within 24-48 hours.”

How Fragile is the Internet?

Today's outage has once again demonstrated how dependent millions of sites and services worldwide are on the infrastructure of a single company. Cloudflare currently carries approximately 20% of global internet traffic, a figure that rises to 35-40% in some countries.

We breathed a sigh of relief; the internet is back. But today was a valuable lesson for all of us: a single wrong line of code can "stop" the world for 102 minutes.

We will continue to update this article as developments occur. You can also share your experiences in the comments section.