addons: Update Prometheus to v2.1.0

* Change service discovery to relabel jobs to align with
rule expressions in upstream examples
* Use a separate service account for prometheus instead
of granting roles to the namespace's default
* Use a separate service account for node-exporter
* Update node-exporter and kube-state-metrics exporters

addons/prometheus/exporters/kube-state-metrics/cluster-role.yaml

@@ -35,4 +35,3 @@ rules:
   resources:
   - horizontalpodautoscalers
   verbs: ["list", "watch"]
-

addons/prometheus/exporters/kube-state-metrics/deployment.yaml

@@ -54,8 +54,8 @@ spec:
           - /pod_nanny
           - --container=kube-state-metrics
           - --cpu=100m
-          - --extra-cpu=1m
-          - --memory=100Mi
-          - --extra-memory=2Mi
+          - --extra-cpu=2m
+          - --memory=150Mi
+          - --extra-memory=30Mi
           - --threshold=5
           - --deployment=kube-state-metrics