helm-chart: Support excluding namespaces

[guikcd]

This is interesting to scan all namespaces except a list of specific ones.
Referenced in #66.

Tested with your two sample deployments, in default namespace, and in mytest namespace (default and kube-system untouched, mytest downscaled):

values.yml:

[...]
namespace:
  inactive_in:
    - kube-system
    - default
[...]

pod describe:

[...]
    Args:
      --interval=60
      --exclude-namespaces=kube-system,default
      --debug
[...]

Logs of kube-downscaler itself:

[...]
2020-01-22 10:48:55,419 DEBUG: Resource flask-v1-tutorial was excluded (either resource itself or namespace None are excluded)
2020-01-22 10:48:55,514 DEBUG: Resource flask-v2-tutorial was excluded (either resource itself or namespace None are excluded)
2020-01-22 10:48:55,522 DEBUG: Resource coredns was excluded (either resource itself or namespace None are excluded)
2020-01-22 10:48:55,522 DEBUG: Resource kube-downscaler was excluded (either resource itself or namespace None are excluded)
2020-01-22 10:48:55,522 DEBUG: Resource tiller-deploy was excluded (either resource itself or namespace None are excluded)
2020-01-22 10:48:55,616 DEBUG: Deployment mytest/flask-v1-tutorial has 1 replicas (original: None, uptime: Mon-Fri 07:00-19:00 US/Eastern)
2020-01-22 10:48:55,616 INFO: Deployment mytest/flask-v1-tutorial within grace period (900s), not scaling down (yet)
2020-01-22 10:48:55,716 DEBUG: Deployment mytest/flask-v2-tutorial has 1 replicas (original: None, uptime: Mon-Fri 07:00-19:00 US/Pacific)
2020-01-22 10:48:55,716 INFO: Deployment mytest/flask-v2-tutorial within grace period (900s), not scaling down (yet)

[coveralls]

Coverage remained the same at 94.942% when pulling 926141b on guikcd:active_inactive_namespaces into 95af6a5 on hjacobs:master.