Monitor device-plugin resources in Kubernetes
We should consider adding support to monitor device-plugin resources in Kubernetes, the foremost candidate for this is GPU's for machine learning: https://kubernetes.io/docs/concepts/cluster-administration/device-plugins/
This would provide visibility and insight into optimizing the quantity of these resources, as well as information if pods are beginning to become unschedulable due to a lack of these resources like GPU 's.
Edited by Joshua Lambert