Helm chart values undocumented

[pierluigilenoci]

What problem are you trying to solve?

There are many completely undocumented configuration possibilities within the chart.

To give some examples:

kubecostToken
kubecostFrontend.tls.enabled
grafana.sidecar.dashboards.enabled (this is more related to Grafana chart but anyway)

Describe the solution you'd like
I'll like to have all the options well documented in the chart README.md so you don't have to guess or go to the chart templates to understand how it works. I want to suggest improving the readability of the whole documentation.

[dwbrown2]

Strong +1 from me. Let's get more documentation added here.

[dwbrown2]

First PR #753 is out. This initial PR adds additional comments for options that exist in values file. We will also pull in other commonly used values. Our subcharts have many flags, so I believe we want to find a balance on the right flags to pull in. We ideally want to focus on commonly used options.

[dwbrown2]

#757 added with more to come

[pierluigilenoci]

networkCosts.trafficLogging is not documented

[dwbrown2]

It's rarely used and therefore documented on the network-allocation docs page.

[pierluigilenoci]

Sparse documentation is not effective for using the chart. IMHO the chart README.md should contain all values and cross-references to in-depth documentation. Regardless of how rarely a configuration is used.

[dwbrown2]

I actually disagree, @pierluigilenoci. If we include all config options, this file will be more than 3k lines long. That will be a bad user experience for most.

My proposal is that we references our subcharts. For example, we can include this for Prometheus.

For more Prometheus configuration options, reference: prometheus/values.yaml

[pierluigilenoci]

Obviously, it makes no sense to document the configuration options for the sub-charts, for that just refer to the official documentation of the sub-charts. I'm talking about the exclusive configurations of kubecost. For example:

• kubecostFrontend.tls.enabled
• networkCosts.trafficLogging

To be able to understand how to configure these options I had to go through the templates of the chart.

[dwbrown2]

Glad we're in agreement on that!

networkCosts.trafficLogging is now included
kubecostFrontend.tls.enabled is technically part of enterprise, so that's why it was not included

I'm not aware of any other missing config options, are you?

[dwbrown2]

Marking as closed as I don't see any others right now we need documented.

[pierluigilenoci]

@dwbrown2
I found one:

kubecostModel:
  utcOffset: "+01:00"

[pierluigilenoci]

Also the new networkPolicy parameters are not documented:

networkPolicy:
  enabled: true
  denyEgress: false
  sameNamespace: false
  namespace: xxx