Cloudflare Blames Outage on Internal Configuration Error

Cloudflare suffered a widespread outage beginning at 11:20 UTC when a permissions change caused ClickHouse queries to generate an oversized 'feature file' for its Bot Management system; the file exceeded size limits, propagated across the network, and triggered repeated 5xx internal server errors for many sites. Cloudflare stopped the bad file generation, manually inserted a known-good file, restarted core proxies, and restored service by 17:06 UTC; the company attributed the incident to an internal configuration error (not a cyberattack) and outlined hardening steps to prevent recurrence.