On November 19, Cloudflare released an analysis report on the large-scale outage that occurred on November 18. The report stated that this outage, the most severe since 2019, was caused by an incorrect configuration file generated after database permission adjustments, which led to an anomaly in the core proxy system and affected multiple products, including CDN, security services, Workers KV, Turnstile, and Access.
The team misjudged it as a DDoS attack and ultimately fixed it by rolling back old files. All services were restored at 01:06 AM Beijing time on the 19th. Cloudflare stated that such downtime is unacceptable and will accelerate system resilience upgrades.
