Given a Node with 10 cpu, if I request a Pod with request 4cpu and limit 6cpu, does that mean the Node has only 4cpu for usage left?
real question here is if I need a Pod with 2 cpu request and 5 cpu limit, as the Node doesn't have 5 cpu left the pod won't be provisioned?
its not clear in the docs
Request are considered during scheduling. Limits are considered while running. Be sure to consider resources reserved for kubernetes as these will not be available to your pods.
The requests are used for scheduling, the limits are that, limits. So, in your example, the node will still have 6 CPU remaining after scheduling your 4 CPU request. It will let your pod use up to 6 CPU, but it will start to limit its performance if it tries to use more than 6, so that it never exceeds 6.
The 2 CPU request with 5 CPU limit can be scheduled on a 2 CPU node, providing that nothing else is running there which requested any CPU.