TechFlow, November 19 — Cloudflare has officially released a detailed post-incident report regarding the global network outage on November 18, 2025. According to the report, the outage began at 11:20 UTC and was not caused by a cyberattack, but rather by a chain reaction triggered by permission changes in the database system. Specifically, altered query behavior in the ClickHouse database caused abnormal growth in Bot Management feature configuration files, exceeding predefined memory allocation limits and resulting in system crashes.
The report thoroughly documents the impact on core services including CDN, Turnstile, and Workers KV during the incident, along with a complete timeline of response actions from 11:28 to 17:06 UTC. The Cloudflare team resolved the issue by halting the propagation of misconfigured settings and rolling back to a stable version.
Cloudflare CEO Matthew Prince acknowledged in the report that this was the company's most severe outage since 2019, and committed to implementing measures such as enhanced configuration validation, additional global feature toggles, and improved error handling mechanisms to prevent recurrence.
