A group of nodes that reside in the same scaling group within a cloud provider, but are also labelled with the same label, e.g. customer: shared
A service that provides compute resources, e.g. AWS, GCP, Azure.
AWS service that groups EC2 instances so that they can be scaled and managed together. Read more about them here.
When we refer to "tainting" a node, we refer to applying a Kubernetes taint to the node that prevents scheduling of pods to that node. We apply the "NoSchedule" effect when we taint a node. More information on tainting of nodes can be found here.
To remove the "NoSchedule" effect from a node. See tainting above.
Extra nodes on top of the needed capacity to run the pods for a node group. This ensures there is capacity for daemonsets to run on each node, as well as some extra capacity to handle a spike in requested resources. Read about slack space here.
The overall node group resource use, based on the resource requests from pods and the allocatable resources from nodes.