Each node in the cluster is assigned a predefined set of common labels. Node labels can be utilized in Pod scheduling decisions by defining node selectors.
Kubedeploy offers this functionality via the
nodeSelector value config option, enabling you to specifically target a particular node or group of nodes.
nodeSelector is undefined.
Define custom nodeSelector
In the above example, Kubernetes scheduler will only consider nodes containing the
karpenter.sh/capacity-type: on-demand label. If no nodes in the cluster have that label Pod will be stuck in
When using cluster automatic scaling software like
Cluster autoscaler, node selectors can be used to influence which types of nodes will be added to the cluster. Please see Best pracices linked below for more info.