bug: Grafana datasource not deploying

[beasteers]

Chart name and version
chart: victoria-metrics-k8s-stack
version: v0.31.0

Describe the bug
Grafana datasource doesn't deploy by default:

    message: 'Helm upgrade failed for release monitoring/victoria-metrics with chart
      victoria-metrics-k8s-stack@0.31.0: YAML parse error on victoria-metrics-k8s-stack/templates/grafana/datasource.yaml:
      error converting YAML to JSON: yaml: line 6: did not find expected key'

It's missing a default value for grafana.sidecar.datasources.label: grafana_datasource

By default it does this:

apiVersion: v1
kind: ConfigMap
metadata:
  name: release-name-victoria-metrics-k8s-stack-grafana-ds
  namespace: monitoring
  labels:
    : "1"

Custom values
Please provide only custom values (excluding default ones):

    grafana:
      enabled: false
      forceDeployDatasource: true
      sidecar:
        datasources:
          label: grafana_datasource  # <<< MISSING

[AndrewChubatiuk]

hey @beasteers
how do you deploy a chart? please share cmd flags you're passing to Helm or spec in case of fluxcd, argocd, etc

[beasteers]

I suspect it's because I disabled the grafana subchart and used forceDeployDatasource.

Just a guess, but perhaps normally it pulls this value (label: grafana_datasource) from the grafana subchart. It might be worth setting a default value here:

helm-charts/charts/victoria-metrics-k8s-stack/templates/grafana/datasource.yaml

Line 14 in ed4ca9a

{{ $.Values.grafana.sidecar.datasources.label }}: "1"

Here's our full Flux HelmRelease.

apiVersion: helm.toolkit.fluxcd.io/v2beta1
kind: HelmRelease
metadata:
  name: victoria-metrics
  namespace: monitoring
spec:
  interval: 15m0s
  releaseName: victoria-metrics
  targetNamespace: monitoring
  chart:
    spec:
      chart: victoria-metrics-k8s-stack
      version: "0.31.0"
      sourceRef:
        kind: HelmRepository
        name: victoria-metrics
        namespace: cluster-config
  install:
    remediation:
      retries: 3
  values:
    defaultDashboards:
      enabled: true

    defaultDatasources:
      victoriametrics:
        datasources:
          - name: Prometheus
            type: prometheus
            isDefault: false

    prometheus-operator-crds:
      enabled: true

    grafana:
      enabled: false
      forceDeployDatasource: true
      sidecar:
        datasources:
          label: grafana_datasource

    prometheus-node-exporter:
      enabled: true
    kube-state-metrics:
      enabled: true
    kubeScheduler:
      enabled: true
    kubelet:
      enabled: true
    kubeApiServer:
      enabled: false #true
    coreDns:
      enabled: false #true
    kubeEtcd:
      enabled: false #true
    kubeControllerManager:
      enabled: false
    kubeDns:
      enabled: false
    kubeProxy:
      enabled: false

    vmsingle:
      spec:
        retentionPeriod: "2w"
        replicaCount: 1
        storage:
          storageClassName: ssd-retain
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 25Gi
        resources:
          requests:
            cpu: 100m
            memory: 600Mi
          limits:
            cpu: 500m
            memory: 600Mi

[AndrewChubatiuk]

copied grafana.sidecar.datasources.label default value from grafana subchart to cover this case, will be available in the next release