Have you considered collecting metrics via another mechanism? I have a puppet manifest that deploys dstat to member nodes, where dstat is configured to send metrics to graphite.
Graphite allows you to build graphs on the fly with very little effort required.
Patrick Debois has an excellent series of blog posts on the subject of metrics and monitoring that begins here:
The upshot is that there are many pieces you can knit together, but if you just want some standard host level metrics, deploying an agent that reports data to graphite is one way to do it that is not very complicated, and yet is quite powerful.