v0.2.1-alpha: Added metrics describing the cpu and mem requests per container . (#35)

These metrics are retrieved using the v1 api directly.
No additional accumulation or computation is done.

The state is maintained per container in a pod in kubernetes.

The metrics are also tagged by the node, where  the containers are residing.
Our use case for these metrics are as follows:

It is useful to accumulate these per node and understand how much
remaining un-requested resourced are available in each node.

This would give an indication of the "utilization" of the cluster.