Capacity decoupling

[etsauer]

This is a refactoring and enhancement to the capacity management script. It now has several components:

• An API module/library for gathering the data and building a json dictionary of all of the data for:
  • The cluster
  • Individual nodes
  • Namespaces, and pods
• Two "front-ends" for fetching the data
  • A CLI (cli.py) that can be run from the command line to print the data, either formatted (like before) or raw json output
  • A web api to serve the json formatted data.

View instructions in capacity-management/README.md for testing.

[sabre1041]

@etsauer hitting the following error on OCP4 and OCP3 environments

new namespace quota
{'cluster': {'quota': {'cpu': '', 'memory': ''}}, 'namespaces': {'esauer-sbx': {'quota': {'cpu': 4, 'memory': 17179869184}}}}
existing namespace quota
{'cluster': {'quota': {'cpu': '', 'memory': ''}}, 'namespaces': {'esauer-sbx': {'quota': {'cpu': 2, 'memory': 12884901888}}}}
Traceback (most recent call last):
  File "./cli.py", line 79, in <module>
    do_print(capacity)
  File "./cli.py", line 8, in do_print
    cluster_memory = capacity["cluster"]["totals"]["memory"]
KeyError: 'totals'

[etsauer]

@sabre1041 fixed.

[sabre1041]

@etsauer looks good. One additional change is to add the application/json Content-Type for the hosted solution. We should be good to merge at that point

[sabre1041]

LGTM