* Add a SourceFetcher, which allows to retrieve new type of data sources.
It allows new collectors to be added.
(examples: Prometheus, Influxdb, Kubernetes...)
* Base Cloudkitty on metrics instead of services and resources
for collecting metrics. This new architecture allows Cloudkitty
to be more agnostic and so to rate containers metrics as the same way
as virtual machines metrics.
* Centralize metrology information in metrics.yml under metrics names.
Task: 6291
Story: 2001501
Change-Id: I00ca080cf05dfc03a3363720f85b79e003eda9be