Particle.news
Download on the App Store
299 ARTICLES
·
yesterday

Cloudflare Restores Service After Global Outage, Blames Bot‑Management Bug

Cloudflare attributes the disruption to a latent bot‑management bug triggered by a routine configuration change, not to an attack.

FILE - Lava lamps are seen through a lobby window at the headquarters of Cloudflare in San Francisco, Wednesday, Aug. 31, 2022. (AP Photo/Eric Risberg, File)
The Cloudflare logo appears on a smartphone screen and as the background on a laptop computer screen in this photo illustration
blank
Cloudflare building

Overview

  • Cloudflare said a fix was implemented around 9:40 a.m. ET and services largely recovered, though it is monitoring for lingering errors and some dashboard and login issues.
  • The failure generated widespread 500 errors that temporarily took major platforms such as X, ChatGPT, Canva, Spotify and Zoom offline for many users, with X reports peaking near 10,000 on Downdetector.
  • Cloudflare’s own control surfaces and tools were affected, including the Dashboard, API, CAPTCHA/Turnstile and WARP, with WARP access in London briefly disabled and later restored.
  • CTO Dane Knecht said a routine change exposed a latent bug underpinning bot mitigation, later detailed as an oversized feature file that exceeded software limits and caused cyclical failures across the network.
  • The incident revived concerns about reliance on a few infrastructure providers, came after recent AWS and Azure outages, and coincided with Cloudflare shares falling about 5% in premarket trading.