How-To: integrate with additional Prometheus server for persistence

Migrated from https://github.com/couchbaselabs/observability/issues/5. 

Demonstrate how to use this in a federated Prometheus set up and with a standalone server just scraping the endpoints.
Also include how this can be used to persist the metric history separately.

Patrick Varley pointed out this would be useful if/when we want to support a "we noticed you've seen this issue before, do you want to fix it like this..." mode.

Also useful would be something like the Lens approach to provide radio buttons to turn on a metrics stack for your cluster: this then deploys the various node exporter and Prometheus operator to track the metrics you want.