Cloudflare: The outage last night was not caused by a cyber attack, but rather a chain reaction triggered by changes in database system permissions.

2025-11-19 01:43:05

Abstract generation in progress

According to Deep Tide TechFlow news, on November 19, Cloudflare officially released a detailed post-incident report on the global network outage that occurred on November 18, 2025. The report indicates that the outage started at 11:20 UTC and was not caused by a cyber attack, but rather by a chain reaction triggered by changes in database system permissions. The specific reason was a change in the query behavior of the ClickHouse database that caused an abnormal increase in the feature profile of the Bot management system, exceeding the preset memory allocation limit, which triggered the system crash.

The report details the impact on several services during the outage, including core CDN, Turnstile, and Workers KV, as well as the complete incident timeline from 11:28 to 17:06. The Cloudflare team resolved the issue by stopping the propagation of the misconfiguration and rolling back to the normal version.

Cloudflare CEO Matthew Prince acknowledged in the report that this is the most severe outage the company has experienced since 2019 and promised to strengthen profile validation, add global feature switches, optimize error handling mechanisms, and take other measures to prevent similar incidents from happening again.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.