Other than throttle_optimize and rebuild_metadata it doesn't look like anything is documented.
I assume you already looked here: http://docs.splunk.com/Documentation/Splunk/6.3.3/Troubleshooting/Aboutmetricslog
It's possible that the undocumented counters are not accurate or even enabled. You should be able to get almost all the troubleshooting data you need from everything that's documented on that page - is there a specific issue you're looking at?
