I have a master node with the etcd role (pic 1). However, etcd metrics are not exposed (pic 2). I am using kube-prometheus-stack instead. How can I monitor the etcd metrics?