+ Services are considered underutilized in GKE when the average usage of + requested cores is less than 80%. +
+ What can I do? ++ Review requested core and limit settings. Check HPA target scaling + settings in hpa.yaml. The recommended value for + targetCPUUtilizationPercentage is 80. +
++ For single pods, there is of course no HPA. But it can also be useful + to think about a single pod out of a larger deployment, then modify + based on HPA requirements. Within a pod, each container has its own + CPU and memory requests and limits. +
+ Definitions ++ A request is a minimum reserved value; a container will never have + less than this amount allocated to it, even if it doesn't actually use + it. Requests are used for determining what nodes to schedule pods on + (bin-packing). The tension here is between not allocating resources we + don't need, and having easy-enough access to enough resources to be + able to function. +
+ + Contact #cost-awareness for information and support. + +