A major outage struck Cloudflare on Tuesday, its most serious since 2019, disrupting major websites including ChatGPT and X. The cloud-infrastructure firm, through CEO Matthew Prince’s blog post, confirmed the root cause was a faulty database query in its Bot Management system — not a cyberattack.
A permissions change triggered an unusual spike in duplicate “feature” rows in the system’s ClickHouse database, causing a configuration file to swell beyond memory limits and crash the core proxy that filters traffic for customers. Legitimate users were blocked as bots, shutting access for users of sites relying on bot-rules. Sites not using that system remained unaffected.
While Cloudflare had recently ramped up tools to manage automated crawlers — especially those training AI models — those new safeguards were not involved in this incident. The company emphasized how global web traffic, with some 20 % routed through its network, remains extremely vulnerable to failures within centralized infrastructure. To prevent a repeat, Cloudflare outlined four immediate remedial steps, acknowledging that internet resilience remains a challenge despite evolving protections.
Read latest news and updates at supernews.pk