How did the internet’s ‘watchman’ Cloudflare fail?

New Delhi, November 19, 2025. Cloudflare, often called the “gatekeeper” of the internet, provides security, speed, and stability to millions of websites, applications, and servers worldwide. Despite this, it recently experienced a massive failure, causing thousands of websites, apps, payment services, and online platforms in multiple countries to suddenly go offline. This incident raised the question: how could such a robust infrastructure collapse in an instant?

It all started yesterday evening. Rahul’s phone began behaving strangely. When he scrolled through X, he saw only a blank screen. He wanted to ask ChatGPT for a dinner recipe, but it showed a “Something went wrong” message. Rahul thought the problem might be with his phone.

Meanwhile, in Mumbai, college student Priya was trying to finish her project on Canva. Suddenly, it stopped working. She turned the Wi-Fi off and on, restarted her phone, but nothing helped. When she called a friend, they were experiencing the same issue.

Gradually, reports emerged that similar problems were occurring worldwide. It turned out there was an issue with Cloudflare’s servers. You could call Cloudflare the internet’s gatekeeper. Website data passes through Cloudflare.

A small permission change activated a latent bug

But how did this happen? Cloudflare released a detailed report of the entire incident in its official blog post. The company admitted this wasn’t a cyber attack, but rather a latent bug in their system that had been there for years, which was activated by a routine change.

• The company’s CTO explained that routine permission changes in the database caused incorrect entries to appear in the bot protection system’s configuration file.
• The file size doubled. When this file was distributed across the global network, thousands of servers crashed because the software couldn’t handle such a large file.
• The file was regenerating every few minutes, which caused the problem to come and go repeatedly. Users were seeing 500 error messages. The issue persisted from 5 PM to 9 PM.

Which services were most affected

Cloudflare provides content delivery network, security, and routing services to over 20% of the world’s websites. The outage affected more than 14 million websites, including X, ChatGPT, WhatsApp, Instagram, Spotify, Canva, Claude AI, Uber, and Zoom. According to one estimate, Cloudflare serves every fifth website in the world.

Cloudflare’s team immediately went into emergency mode, rolling back the update, refreshing server caches, and rerouting traffic between data centers. Service was restored within hours, but this incident made it clear that even the world’s strongest tech companies can fall victim to a ‘single point of failure’.

This failure also showed how much of the global internet operates on dependency—a single company’s mistake can shake the entire digital ecosystem. This event emphasizes the need for more secure, decentralized, and multi-cloud infrastructure in the future.