Your integration talks to a single hostname. Routing to the nearest healthy node happens underneath, and fails over automatically if one goes down.
Capacity expands automatically ahead of announced traffic spikes. Enterprise plans can request dedicated node capacity in any existing region — see pricing.
Anycast handles the first hop; health checks handle the rest.
A single IP is announced from every PoP. BGP routes each client to the topologically nearest one without any client-side configuration.
Every node is probed every 5 seconds. A node failing 3 consecutive checks is pulled from rotation before it can serve a broken response.
If a region goes dark, traffic reroutes to the next-nearest healthy PoP inside the current TCP session where possible, or on next connect.