Use Node Allocatable Resources instead of raw Node Capacity

[hjacobs]

https://github.com/kubernetes/community/blob/master/contributors/design-proposals/node-allocatable.md

[hjacobs]

Hmm, in our Zalando AWS setup the values are currently identical:

kubectl get nodes ip-172-31-1-1.eu-central-1.compute.internal -o yaml
...
 allocatable:
    alpha.kubernetes.io/nvidia-gpu: "0"
    cpu: "2"
    memory: 15661512Ki
    pods: "110"
  capacity:
    alpha.kubernetes.io/nvidia-gpu: "0"
    cpu: "2"
    memory: 15661512Ki
    pods: "110"
...

[hjacobs]

The necessary flags need to be passed to kubelet to get different values for allocatable and capacity:

• --kube-reserved mapStringString
  A set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G) pairs that describe resources reserved for kubernetes system components. Currently only cpu and memory are supported. See http://kubernetes.io/docs/user-guide/compute-resources for more detail. [default=none]
• --system-reserved mapStringString
  A set of ResourceName=ResourceQuantity (e.g. cpu=200m,memory=150G) pairs that describe resources reserved for non-kubernetes components. Currently only cpu and memory are supported. See http://kubernetes.io/docs/user-guide/compute-resources for more detail. [default=none]

[hjacobs]

BTW: the contrib/cluster-autoscaler also does not honor allocatable resources right now 😞