Basically, if a resource knows how it starts up, then it likely knows which process it creates in the process table (or that process can be derived from reading a pid file, etc). So, if we support this resource-process association, then we can monitor that access process.

Then, some low-level / native process metrics we would want to collect might be:

* Availability
* Cpu System Time (amd per minute)
* Cpu Total Time (amd per minute)
* Cpu User Time (amd per minute)
* Cpu Usage
* Memory Size
* Open File Descriptors
* Virtual Memory Size