Cloudflare CEO Matthew Prince confirmed that a huge global internet outage that disrupted millions of users worldwide was caused by an internal malfunction, not a cyberattack. The outage impacted many other platforms, such as X, ChatGPT, Canva, Discord, and many other websites and apps too, leaving users and businesses unable to access services for several hours.
What went wrong?
According to Prince, the problem began when Cloudflare updated permissions on a ClickHouse database cluster. Although the update was made to improve access to data across the network, but a defective query accidentally pulled in far more information than the system could handle. This caused a key “feature file” used by Cloudflare’s Bot Management system to grow far larger than its design limit, forming turbulence across the network.
The feature file is generally refreshed and shared across Cloudflare’s global network within every five minutes. When the file unexpectedly doubled in size, it exceeded software limits, causing the routing systems at the network edge to crash. This ongoing cycle of crashes and temporary fixes stayed for about three hours, starting around 11:20 UTC, leading to service disruptions across the network.
Initially misread as a cyberattack
At first, Cloudflare engineers presumed the outage was the result of a large-scale DDoS attack, due to the sudden service failures. However, after a breif investigation, the company assured that the cause was an internal error. Prince stated that no hacking or malicious activity was involved in the incident.
The company confirmed that the outage was fully resolved by 17:06 UTC. Prince described the incident as the most serious service disruption Cloudflare has experienced since 2019 and acknowledged the inconvenience caused to users and businesses worldwide
Prince emphasized that the company is committed to learning from this incident and making its network more reliable for users, partners, and businesses around the world. As part of this effort, Cloudflare is also reviewing its key systems to see where problems could occur and finding ways to make them stronger and more dependable.
Doonited Affiliated: Syndicate News Hunt
This report has been published as part of an auto-generated syndicated wire feed. Except for the headline, the content has not been modified or edited by Doonited
