This incident is now resolved. Over the past several days, traffic from some clusters to Cloudflare and Cloudflare-hosted services saw packet loss, slow downloads, and intermittent connection failures. We traced this to congestion on an upstream network path between our facility and Cloudflare, rather than an issue inside our datacenter or on the Cloudflare platform itself. In coordination with Cloudflare and the network provider, traffic was rerouted around the affected segment and performance has recovered. All components are operational, and we are continuing to monitor to ensure it holds. We appreciate your patience while we worked through this.

We have observed a consistent increase in performance and a decrease in packet loss for traffic transiting between one of our facilities and Cloudflare. We will continue to monitor and confirm a resolution.

We are continuing to track an issue in routing between all of our ISPs and Cloudflare that materializes as intermittent packet loss and decreased throughput to Cloudflare services and some sites behind Cloudflare's WAF/CDN for some customer clusters. This issue has been escalated to Cloudflare for resolution.