*Coverage analysis tooling

*Coverage nightly run is failing:
	* Coverage collection from the coverage module stops before api call to do so is run

*Changes to existing coverage extension:
	* Drop backdoor port and move to rpc
		* put in common rpc layer
		* Add multihost support
	* Common extension/controller code in oslo?

*Port extension to other projects?
	* Common code in oslo or somewhere else?
		* Common wsgi code so the api semantics are the same or just an extension ported to every project
		* Move coverage collection to be only rpc and remove eventlet backdoor part (which is kind of weird)

*Additional tooling around coverage analysis
	* Take the full coverage report and automate interpretation of results. (Right now it's manual by inspection)
		* API coverage report
			* Nova API coverage report is generated from this tool.
			https://docs.google.com/spreadsheet/ccc?key=0AmYuZ6T4IJETdEVNTWlYVUVOWURmOERSZ0VGc1BBQWc
			tool: https://review.openstack.org/#/c/25882/ (But abandoned. it needs to be refined.)
			
Current coverage extension:
https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/contrib/coverage_ext.py
http://git.openstack.org/cgit/openstack/nova/tree/nova/api/openstack/compute/contrib/coverage_ext.py :)

Should coverage collection be enabled for every tempest run?


Actions:
	*     Clark and Matt to look at coverage failures
	* Add coverage run in experimental queue
	*      Add Ability to configure costum wrapper for every python service process  @ devstack (afazekas)
	* Graphite meters
		* tempest runs per project / service
		* unit tests
	* refresh  https://review.openstack.org/#/c/25882/  for Nova v2 API(cyeoh, masayuki)