Global Cloudflare Disruption on June 12, 2025
On June 12, 2025, Cloudflare outage experienced a severe service outage that disrupted several of its critical products and affected customers worldwide. The outage, which lasted for 2 hours and 28 minutes, impacted many of Cloudflare’s widely used services, including Workers KV, WARP, Access, Gateway, Images, Stream, Workers AI, Turnstile, Challenges, AutoRAG, Zaraz, and sections of the Cloudflare Dashboard.
The incident was traced back to a failure in the underlying storage system used by Workers KV, a service crucial for storing configuration data, authentication information, and asset delivery for multiple Cloudflare products. This storage layer partly depends on a third-party cloud provider, which suffered its outage, directly affecting Cloudflare’s operations. While no customer data was lost and core security systems like Magic Transit, DNS, and WAF continued operating normally, many services experienced high failure rates, widespread errors, and significant disruptions in user access and functionality.
Services Impacted Across the Cloudflare Ecosystem
The failure of Workers KV had cascading effects across numerous Cloudflare outage services. Access, which depends heavily on Workers KV for identity and policy configurations, saw 100% login failures during the outage. Services like WARP and Gateway, which rely on Access for device registration and authentication, also faced severe disruptions. New WARP connections were blocked, and Gateway’s identity-based DNS filtering and proxy services failed to function for many users.
The Cloudflare outage Dashboard became mostly inaccessible as login systems tied to Turnstile, Access, and Workers KV went offline. Similarly, the Turnstile and Challenges platforms experienced significant failure rates due to the inability to validate user tokens. Browser Isolation sessions were interrupted as they depended on the impaired Gateway service.
Cloudflare’s image and video services also took a hit. Image uploads failed at the peak of the incident, while video services under Stream saw error rates exceeding 90%, with Stream Live unable to deliver content. Other services like Realtime, Workers AI, Pages, AutoRAG, Durable Objects, D1, Queues, AI Gateway, and CDN, Workers for Platforms, Browser Rendering, and Zaraz faced varying degrees of service failures, primarily tied to their reliance on Workers KV’s storage infrastructure.
Cloudflare’s Response and Future Measures
Cloudflare outage engineers quickly initiated recovery efforts once the problem was identified. The first signs of trouble appeared around 17:52 UTC when the WARP team noticed new device registrations failing. The incident was escalated rapidly, and by 20:28 UTC, services began stabilizing as the affected storage systems recovered.
In response to the incident, Cloudflare outage has committed to accelerating several resilience improvements. Key efforts include removing single points of failure within Workers KV’s storage layer, introducing alternate data stores for critical services, and deploying tools to allow more controlled recovery during similar outages in the future. Additionally, Cloudflare plans to strengthen service independence to ensure failures in one area do not cascade across multiple services.
Cloudflare outage has publicly apologized for the disruption, recognizing the trust organizations place in its infrastructure to support websites, applications, and networks globally. The company reassured customers of its ongoing dedication to improving service stability and avoiding similar outages in the future.
Visit more of our news! Mr. Business Magazine.
