Cloudflare service outage June 12, 2025

On June 12, 2025, Cloudflare experienced a 2h28m global outage impacting Workers KV and numerous dependent services (including Access, Gateway, WARP, Dashboard, Turnstile, Workers AI, Durable Objects, D1, Stream, and more) due to a failure in third-party-backed storage infrastructure used as KV’s central store; some cache-served traffic continued but many components failed closed. The event was not an attack and no data was lost; the company escalated to P0, applied mitigations (e.g., kill switches, dependency shedding, prioritizing critical services), and services recovered around 20:23–20:28 UTC as the storage backend returned. Cloudflare is accelerating efforts to remove single points of failure, increase KV redundancy (including migration to R2), and add tooling for progressive namespace re-enablement to reduce blast radius.