Add kube-proxy health check

[dvdthms]

We had noticed that there wasn't a check for the kube-proxy healthz endpoint in the cfn-signal service.

We hit an issue where the cfn-signal was firing, but the proxy was not healthy (it could not talk to the API-server). We've added this check to make sure the proxy is healthy before allowing the cfn-signal to fire.

Out of interest -- was there a reason this check wasn't included in the command before?

The port used for the kube-proxy healthz can be seen here: https://github.com/kubernetes/kubernetes/blob/v1.7.0/pkg/master/ports/ports.go#L41-L43

[danielfm]

This part of the code comes from the early days of kube-aws, it's not a surprise some things may have been overlooked back then.

/lgtm

[codecov-io]

Codecov Report

Merging #1004 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1004   +/-   ##
=======================================
  Coverage   34.88%   34.88%           
=======================================
  Files          59       59           
  Lines        4133     4133           
=======================================
  Hits         1442     1442           
  Misses       2532     2532           
  Partials      159      159

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 976e946...84d18b4. Read the comment docs.

[mumoshu]

was there a reason this check wasn't included in the command before?

No. Thanks for the good catch!