Sure. Generally we'd just like to ingest metrics for certain databases, loadbalancers, and other resources that we care about. Currently we can get them in Cloudwatch itself, or another collection system.
However it's nice to have them in Elasticsearch so that we can visualize them along side other things (like logs, custom metrics, etc). It also lets us write more specific alerts (i.e. when some custom metric and some database metric are out of the ordinary, simultaneously).
The rabbitmq module is nice because it let's you pick resources based on a regex, which might be nice here too (all prod databases,
prod-*, where we don't care about dev/test/etc).
I'd be happy to add more detail in any particular area if you have follow ups.