I am running a jupyterhub on k8s on a supercomputer (1node, 8cpu,16Gi).
I added a cpu and memory request/limit to the config.yaml as shown above.But the number of singleuser is over 40 servers.why?8/0.3=26.6,the number of singleuser no more than 27 servers, right?
limit controls the maximum resources the pod is allowed to use, if the pod uses less than that then another pod can use the resource. guarantee means those resources are reserved for that pod only.