Cybersecurity

November 18, 2025 — Cloudflare experienced a significant outage that affected many major websites and services. This incident was caused by a bug in the Bot Management feature, which led to a crash in the software handling traffic. The outage began around 11:20 UTC and lasted for several hours, impacting approximately one-third of the world’s top websites, including X (formerly Twitter) and ChatGPT.

The outage was actually caused by Cloudflare’s own system. The cause involved a “feature file,” a file that tells Cloudflare software what kind of malicious behavior bots may exhibit. Such behavior is constantly changing in a never-ending cat-and-mouse game between threat actors and suppliers such as Cloudflare. The Bot Management system receives this feature file, which contained over 200 features.

The Cloudflare software responded to the unusual over 200 feature file resulted in the system “panicking.” This abundance of features caused erroneous change in the permissions of a database system. Multiple entries appeared in the database for the now infamous feature file, with all the consequences that entailed.

Prior to this, Cloudflare had a notable incident on October 3, 2025, involving a 47-minute DNS resolver failure that made millions of websites unreachable. As well as , a January incident caused by a cache purge bug leading to cascading failures, and a September outage from a dashboard bug that overwhelmed an API service.These issues were part of a year of various challenges for Cloudflare, which included a significant data breach involving third-party vendors in early 2025.

What happened in January

January 11, 2025: Cache Purge Bug 

A latent bug in the cache purge system was triggered, causing cascading failures across the network. The Content Delivery Network (CDN) cache hit rates plummeted, sending a flood of requests to origin servers which many could not handle, resulting in widespread errors and a two-hour disruption.

Ongoing DDoS Attacks

 Cloudflare filtered a massive DDoS (Distributed Denial-of-Service) attack in January 2025 with a peak traffic volume of 5.6 Terabits per second. This was an external issue where attackers used a botnet of approximately 13,000 IoT devices, part of a growing trend of larger and more complex attacks seen throughout 2025.

Third-Party Disruptions 

Cloudflare’s network also experienced some impact from external issues, such as an internal fault in the AAE-1 submarine cable near Qatar on January 2 that caused increased latency in Pakistan, and power outages in Angola and Ireland (due to severe storms) that disrupted local connectivity in those regions. 

What happened in October

•  A bug in the DNS resolver software, which was part of an automatic update that lacked proper testing, caused a complete DNS resolution failure.
• The approximately 12 million websites using Cloudflare’s DNS service became unreachable for 47 minutes.
• A bug in the DNS resolver software slipped through automatic updates because there was no proper testing, such as a canary test, before the deployment. 

What happened in November

• Major platforms like Shopify, Zoom, Canva, and Coinbase were among those that experienced downtime.
• Thousands of users reported issues accessing these services, with many encountering error messages.
• It was estimated that about 20% of all websites were affected at the peak of the outage.

Response and Recovery

Cloudflare acknowledged the issue and apologized for the disruption. They stated that the problem was not due to a cyberattack but rather an internal technical failure. Engineers worked to resolve the issue by halting the faulty feature file generation and restoring normal traffic flow by around 14:30 UTC.

Future Measures

In response to this incident, Cloudflare has committed to improving their internal processes. They plan to enhance validation for configuration files and implement more robust fail-safes to prevent similar outages in the future. This incident has highlighted the reliance on Cloudflare’s services, as many organizations depend on them for web security and performance.

X

X suffered its third major outage in just four days, with thousands of users across the globe reporting problems on Thursday.

Data from Downdetector showed a surge in reports, with users in the UK, Europe, the US and beyond saying their timelines and profiles would not load.

Many affected users received the error message: “Something went wrong. Try reloading.” Most issues were reported on the mobile app, though the desktop site was also affected.

On Tuesday, X was offline for around an hour, leaving millions of users unable to access the platform, while another disruption occurred shortly before that.

The latest outage follows widespread internet problems caused by a major Cloudflare failure, which temporarily affected several high-profile sites and services due to a global network configuration error.

As of Thursday afternoon, some users were still reporting intermittent access problems, though reports suggested service was gradually stabilizing.