### Description of problem ###

I'm not sure if this is bug.

Per default the prometheus serviceaccount gets the view clusterrole, which is not able to view/list/watch nodes in the cluster. However prometheus tries to acquire node objects in the cluster and creates following error messages:
~~~
 k8s.io/kube-state-metrics/collectors/node.go:130: Failed to list *v1.Node: nodes is forbidden: User "system:serviceaccount:openshift-metrics:default" cannot list nodes at the cluster scope: User "system:serviceaccount:openshift-metrics:default" cannot list all nodes in the cluster
~~~

After adding a cluster role with the right to view nodes the mistake is gone.

If prometheus is deployed with ansible.
It's a permanent mistake, because prometheus (with the default serviceaccount) continuously tries to view nodes.

### Version-Release number ###

 openshift-ansible-3.10.21-1.git.0.6446011.el7.noarch