global.fleet.clusterLabels not rendered in plain list items

[gervaso]

Is there an existing issue for this?

• [X] I have searched the existing issues

Current Behavior

I can use the Grafana labs Helm Chart as an example and I think this bug is related to #363 and #402

I defined a clusterlabel myproduct/hostname and in one particular case it doesn't get rendered

defaultNamespace: grafana
helm:
  repo: https://grafana.github.io/helm-charts
  chart: grafana
  releaseName: grafana
  version: 6.32.12
  values:
    serviceMonitor:
      enabled: true
      path: /metrics
    ingress:
      enabled: true
      ingressClassName: nginx
      annotations:
        # cert-manager.io/cluster-issuer: generic-cluster-issuer
        ingress.kubernetes.io/proxy-body-size: "0"
        # ingress.kubernetes.io/ssl-redirect: "true"
      labels: {}
      path: /grafana
      pathType: Prefix
      hosts:
        - global.fleet.clusterLabels.myproduct/hostname
      # tls:
      # - secretName: grafana-tls
      #   hosts:
      #     - global.fleet.clusterLabels.myproduct/hostname
    persistence:
      type: pvc
      enabled: true
      storageClassName: openebs-zfspv
      accessModes:
        - ReadWriteOnce
      size: 1Gi

The cluster labels appear correctly in the rendered values, but not in the "hosts" list

> helm get values -n grafana grafana
USER-SUPPLIED VALUES:
global:
  fleet:
    clusterLabels:
      brainkin-fleet: enabled
      myproduct/hostname: myhostname.int.company.tld
      management.cattle.io/cluster-display-name: myproduct-dev
      management.cattle.io/cluster-name: c-m-xvxjw2sg
      provider.cattle.io: k3s
ingress:
  annotations:
    ingress.kubernetes.io/proxy-body-size: "0"
  enabled: true
  hosts:
  - global.fleet.clusterLabels.myproduct/hostname
  ingressClassName: nginx
  labels: {}
  path: /grafana
  pathType: Prefix
persistence:
  accessModes:
  - ReadWriteOnce
  enabled: true
  size: 1Gi
  storageClassName: openebs-zfspv
  type: pvc
serviceMonitor:
  enabled: true
  path: /metrics

Expected Behavior

I would expect it to correctly render the clusterlabels value of "myhostname.int.company.tld" as it happens with other charts

Steps To Reproduce

No response

Environment

- Architecture: amd64
- Fleet Version: 0.3.9
- Cluster:
  - Provider: k3s (the the same happens with vanilla k8s or rke)
  - Options: the cluster is joined to a rancher server running version 2.6.6
  - Kubernetes Version: 1.23.8

Logs

No response

Anything else?

No response