For readers this is how we use Cloudflare is why this matters: WNAM API <> Cloudflare <> Supabase We use Cloudflare to protect and accelerate your requests to our database and API. Requests that were previously made within LA by our API to our database at us-west were carried and responded to within LAX, this meant low latencies and fast responses. Cloudflare is now responding to those requests made in LA from: - Portland - Dallas - Denver - Dublin and more. This means our API is now responding much slower as traffic originates in LA, travels to Cloudflare in LAX, reaches us-west and has to travel back to Cloudflare which then serves it from other states and another continent. We're pursuing multiple remediation paths including removing Cloudflare entirely to restore fast responses via our peering with AWS directly.

The friendly people on Discord said official channels are the best way to get support. I really don't like making a public stink on X but the response from your support makes it seem like latency is not a priority for Cloudflare. "Cloudflare uses anycast routing, and our traffic engineering can deliberately steer traffic to a slightly farther data center when the closest location is congested or otherwise unstable, including during and after the network event you linked. In those situations reliability is prioritized over geographic closeness." I understand there may be reasons for entirely turning off a region but your status page says everything is normal at LAX - which it clearly is not. Attached is a screenshot from your dashboard showing how serious this problem is. That said I have responded to your email with the requested proof - but surely you guys have awareness over your network and don't need users to do debugging for you?

more information about account credits will be available soon !!