fix(): values path for CRDs when running upgrade job

[termin-natetor]

[kube-prometheus-stack] Fix incorrect values path in CRD upgrade job template

What this PR does / why we need it:

When rolling out a major upgrade from 85.0->86.0, I noticed that even when I tried to disable the upgrade Job (we use Argo), it was still being created.

I poked around and it looks like the Helm template is using the wrong key:
.Values.upgradeJob.* instead of .Values.crds.upgradeJob.*

Special notes for your reviewer

Checklist

• DCO signed
• Chart Version bumped
• Title of the PR starts with chart name (e.g. [prometheus-couchdb-exporter])

[termin-natetor]

Per docs, at'ing the maintainers:
@andrewgkew
@gianrubio
@gkarthiks
@jkroepke
@Xtigyro
@QuentinBisson

[jkroepke]

Per docs, at'ing the maintainers:

which docs telling you to at'ing the maintainers?

[jkroepke]

Mention that the jobs lives in the sub chart crd. Values from the main. chart to the sub chart needs to be prefixed by the sub chart name itself. Any values under .Values.crds will be passed as .Values to the sub chart. .Values.upgradeJob.* is the right key inside the template.

[termin-natetor]

Hiya,

Per docs, at'ing the maintainers:

which docs telling you to at'ing the maintainers?
When you open up a PR, it is part of the template - is that generated by github and not your repo?

In any case thanks for responding. I suspected that would be the case but that's not the behavior I see in my chart, is there a test for that? I had to set values to this in our ArgoCD wrapper chart to get the behavior I expected (no upgrade job):

kube-prometheus-stack:
  upgradeJob:
    enabled: false
  crds:
    upgradeJob:
      # explicitly disable the upgrade job so Argo can manage it
      enabled: false

[jkroepke]

This should be sufficent:

kube-prometheus-stack:
  crds:
    upgradeJob:
      # explicitly disable the upgrade job so Argo can manage it
      enabled: false

Ref:

helm-charts/charts/kube-prometheus-stack/values.yaml

Line 40 in 1332ee7

enabled: false

The default is false anyways.

[termin-natetor]

Thats not the behavior I got, here is a little excerpt from Argo:

As soon as I added in the "reduntant" flag, as I showed, it worked - i.e. the job was not created and the upgrade was successful.

[jkroepke]

The chart itself works as expected:

• crds.upgradeJob.enabled=false -> No Upgrade Job
• crds.upgradeJob.enabled=true -> Upgrade Job present.

jok@joe-air ~ % helm template kube-prometheus-stack -n monitoring oci://ghcr.io/prometheus-community/charts/kube-prometheus-stack --set crds.upgradeJob.enabled=false | grep kube-prometheus-stack-crds-upgrade
Pulled: ghcr.io/prometheus-community/charts/kube-prometheus-stack:86.2.0
Digest: sha256:2182a9de8cd44d7a7ac98aa04743de4335766376cc896b9a6e099c61e81fd2d3
jok@joe-air ~ % helm template kube-prometheus-stack -n monitoring oci://ghcr.io/prometheus-community/charts/kube-prometheus-stack --set crds.upgradeJob.enabled=true | grep kube-prometheus-stack-crds-upgrade
Pulled: ghcr.io/prometheus-community/charts/kube-prometheus-stack:86.2.0
Digest: sha256:2182a9de8cd44d7a7ac98aa04743de4335766376cc896b9a6e099c61e81fd2d3
  name: kube-prometheus-stack-crds-upgrade
  name: kube-prometheus-stack-crds-upgrade
  name: kube-prometheus-stack-crds-upgrade
  name: kube-prometheus-stack-crds-upgrade
    name: kube-prometheus-stack-crds-upgrade
  name: kube-prometheus-stack-crds-upgrade
  name: kube-prometheus-stack-crds-upgrade
      serviceAccountName: kube-prometheus-stack-crds-upgrade
            name: kube-prometheus-stack-crds-upgrade

---

With this PR, the value to enable the crds-upgrade Job would be --set crds.crds.upgradeJob.enabled=true