How does the Kubernetes apiserver handle clients http requests?

Hi all,

When a client (e.g scheduler, kubelets, kube-proxy, controllers, kubectl) sends a request to the apiserver, how does the latter handle that request ? I mean does it spin up a go-routine for each incoming request or does it have some queue management internally and a pool of threads/go-routines ? And can a high latency client (e.g a kubelet a very far node on the internet) affect the processing time of other requests (maybe because other requests must wait until the response is sent back which takes a longtime or something that) ?

Thanks,