(MENAFN- Khaleej Times)

Khaleej Times' website and multiple online sites, including US President Donald Trump's Truth Social network and online graphic design platform Canva, were briefly down on Friday following a widespread outage at US-based internet infrastructure firm Cloudflare.

It was the second time in two months that the US-based internet infrastructure - that powers more than a fifth of global web traffic - suffered from“internal service degradation”, resulting in connectivity problems across multiple websites, apps and platforms worldwide.

Recommended For You

Cloudflare provides key internet infrastructure, including services helping websites load faster and shielding them from cyberattacks to enable them to stay accessible during high traffic. Barely a month ago, however, an outage at Cloudflare also prevented thousands of users from accessing major online platforms, including X and ChatGPT.

Stay up to date with the latest news. Follow KT on WhatsApp Channels.

What happened

Khaleej Times spoke to Dubai-based technologist and cybersecurity expert Rayad Kamal Ayub to break down what happened and explain how this outage can be prevented in the future.

Ayub noted Cloudflare's maintenance work on Friday (December 5) and its inability to transfer traffic caused widespread disruptions.

“Cloudflare is described as an 'Internet super node', handling a significant portion of global web traffic (about 20.5 per cent). When Cloudflare experiences issues, the effects ripple across the entire digital world,” explained Rayad, who is also managing director of UAE-based Rayad Group.

Simply put,“Cloudflare acts as a middle layer between users and websites. It helps websites load faster while protecting them from cyberattacks. With many companies and platforms relying on Cloudflare, a problem in its system - even if they are unrelated - can result in websites to crash.”

Failure in automation

“The outage in November 2025 was not due to a cyberattack - and today it was maintenance - but rather a failure in automation,” Ayub added, noting:“A routine permissions change in the configuration management system led to duplicate entries in a database. This caused a configuration file for the bot management system to double in size unexpectedly.

The oversized configuration file was pushed to Cloudflare's global network. Systems with hard memory limits crashed when they received the file, resulting in widespread HTTP 5X or server errors.

Ayub said services that rely on real-time edge routing - like ChatGPT and Coinbase - failed instantly because they have no fallback capacity. Unlike static websites, these platforms cannot cache (store) dynamic content, so any disruption at the edge gateway leads to total service failure.”

How to avoid the next crash

Ayub underscored“the outage demonstrated that the internet's infrastructure is highly centralised and, therefore, fragile. A simple, non-malicious bug in a configuration file can temporarily cut off access to essential AI and financial tools, raising concerns about resilience and access as fundamental issues.”

The two recent incidents exposed the systemic risk of relying on a single provider. Ayub, highlighted the need for:

• Multi-provider, multi-CDN (content delivery network), and multi-DNS (domain name system) strategies for resilience.

• Better isolation between control and data planes.

• Safety checks in automation to prevent oversized files from being deployed.

• Improved fallback mechanisms to revert to last known good configurations instead of crashing.

MENAFN05122025000049011007ID1110443463