Under section 11 "memory reclaim" the author states: >"But memory request is actually used only internally e.g. to calculate oom_adj and QoS." Could someone e

Under section 11 "memory reclaim" the author states: >"But memory request is actually used only internally e.g. to calculate oom_adj and QoS." Could someone elaborate on this? Doesn't Kubernetes also use this to decide if a container could be scheduled on this node in the first place? Or am I missing some other obvious point the author is trying to make?