Merge pull request docker#7678 from david-yu/patch-11

Add detail on why to use Prometheus on worker nodes

Author: ddeyo

ee/ucp/admin/configure/collect-cluster-metrics.md

@@ -64,6 +64,8 @@ These are application-level metrics derived from the container platform itself.
 
 ## Deploy Prometheus on worker nodes
 
+Universal Control Plane deploys Prometheus by default on the manager nodes to provide a built-in metrics backend. For cluster sizes over 100 nodes or for use cases where scraping metrics from the Prometheus instances are needed, we recommend that you deploy Prometheus on dedicated worker nodes in the cluster.
+
 To deploy Prometheus on worker nodes in a cluster:
 
 1. Begin by sourcing an admin bundle.
@@ -76,7 +78,7 @@ To deploy Prometheus on worker nodes in a cluster:
     ucp-metrics-hvkr7   3/3       Running   0          4h        192.168.80.66   3a724a-0
     ```
 
-3. Add a Kubernetes node label to one or more workers.  Here we add a label with key "ucp-metrics" and value "".
+3. Add a Kubernetes node label to one or more workers.  Here we add a label with key "ucp-metrics" and value "" to a node with name "3a724a-1".
 
     ```
     $ kubectl label node 3a724a-1 ucp-metrics=