Issue-581 Make metrics exposable based on helm variables

[joshsouza]

Resolves: #581

[codecov]

Codecov Report

Attention: 6 lines in your changes are missing coverage. Please review.

Comparison is base (88b9307) 85.12% compared to head (794565c) 85.91%.
Report is 1 commits behind head on master.

Files	Patch %	Lines
controllers/zookeepercluster_controller.go	70.00%	3 Missing and 3 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #582      +/-   ##
==========================================
+ Coverage   85.12%   85.91%   +0.79%     
==========================================
  Files          12       12              
  Lines        1613     1633      +20     
==========================================
+ Hits         1373     1403      +30     
+ Misses        155      145      -10     
  Partials       85       85              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jkhalack]

Has there been any testing done to prove that the modified chart allows exposing metrics for Prometheus?

[jkhalack]

Does it have to be -metrics-bind-address or --metrics-bind-address?

[joshsouza]

Based on:

zookeeper-operator/main.go

Line 65 in 074d8b0

flag.StringVar(&metricsAddr, "metrics-bind-address", "127.0.0.1:6000", "The address the metric endpoint binds to.")

And flag's documentation: https://pkg.go.dev/flag#hdr-Command_line_flag_syntax
My understanding is that it can be either. If there is a preference for double dashes, that's easy to adjust, but as -disableFinalizer below is a single dash, I went with single dash for consistency.

[joshsouza]

Has there been any testing done to prove that the modified chart allows exposing metrics for Prometheus?

I utilized helm template when writing the code to ensure the appropriate adjustments to the k8s resources, and deploying it to a local kind cluster before/after with a second curl pod deployed to test connectivity validates that without these changes I am unable to curl <pod_ip>:6000/metrics (connection refused) beforehand, and with --set metricsBindAddress=0.0.0.0 in my helm update/install parameters, I am able to successfully reach the metrics endpoint from my curl pod afterwards.

While I didn't explicitly test that Prometheus itself was working, this should be a sufficient test to verify that the changes are functioning as intended, and should allow metrics gathering by other pods in the cluster.

[anishakj]

LGTM