Cloudflare has confirmed that the significant service outage experienced yesterday was not attributed to any security incident, assuring users that their data remains intact. The outage commenced at 17:52 UTC, primarily affecting the Workers KV (Key-Value) system, which is essential for many of Cloudflare’s services, leading to widespread service disruptions across various edge computing and AI platforms.
According to a post-mortem published by Cloudflare, the outage persisted for nearly 2.5 hours. The company identified the outage’s root cause as a failure within the underlying storage infrastructure of the Workers KV service, which significantly relies on a third-party cloud provider. This infrastructure failure directly impacted the availability of Cloudflare’s KV service, subsequently affecting a range of other services.
The Workers KV system, a globally distributed and consistent key-value store, saw an alarming 90.22% failure rate due to backend storage unavailability. Critical failures were reported across multiple services, including Access and WARP, both of which handle identity-based authentication and session management. Additionally, functionality in tools such as the Dashboard and Turnstile was significantly impaired, with login and verification failures becoming commonplace.
In response to this incident, Cloudflare has pledged to expedite efforts to enhance system resilience by reducing its reliance on a single third-party provider for Workers KV backend storage. The company plans to gradually transition the KV’s central store to its own R2 object storage, alongside developing new cross-service safeguards and recovery tools to mitigate impacts during future storage outages. These changes aim to prevent the kind of cascading failures seen during this latest event. For further details, users can refer to the information on the Cloudflare service outage.