Dashboards

Dashboards is the core feature of Grafana and of course something that you can manage through the operator.

To view the entire configuration that you can do within dashboards, look at our API documentation.

Dashboard managment

You can configure dashboards as code in many different ways.

json
gzipJson
URL
Jsonnet

Json

A pure JSON representation of your Grafana dashboard. Normally you would create your dashboard manually within Grafana, when you have come up with how you want the dashboard to look like, you export it as JSON, grab the JSON using the export function in grafana and put inside the GrafanaDashboard CR.

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-sample
spec:
  resyncPeriod: 30s
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  json: >
    {
      "id": null,
      "title": "Simple Dashboard",
      "tags": [],
      "style": "dark",
      "timezone": "browser",
      "editable": true,
      "hideControls": false,
      "graphTooltip": 1,
      "panels": [],
      "time": {
        "from": "now-6h",
        "to": "now"
      },
      "timepicker": {
        "time_options": [],
        "refresh_intervals": []
      },
      "templating": {
        "list": []
      },
      "annotations": {
        "list": []
      },
      "refresh": "5s",
      "schemaVersion": 17,
      "version": 0,
      "links": []
    }

gzipJson

It’s just like JSON but instead of adding pure JSON to the dashboard CR you add a gzipped representation. This allows you to do really big dashboards while not hitting the etcd maximum request size of 1,5 MiB.

Assuming a dashboard is already saved in dashboard.json, the steps below describe how you can prepare a CR.

cat dashboard.json | gzip | base64 -w0

Take the output and put it in your GrafanaDashboard CR, for example:

---
apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-gzipped
spec:
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  gzipJson: |-
    H4sIAAAAAAAAA4WQQU/DMAyF7/0VVc9MggMgcYV/AOKC0OQubmM1jSPH28Sm/XfSNJ1WcaA3f+/l+dXnqk5fQ6Z5qf3eubt5VlKHCTXvNAaH9RtE2zKI2fQnCgFNsxihj8n39V3mqD/zQwMyXE004ol95q3wMaIsEhpSaPMTlT0WasngK3sVdlN6By4uUi8Q7AezUwpJeig4gEe3ajItTfM5T5l0wuNUwfNx82RLg9nLhTeZXW4iAu2GVHcVNPEtByX2tyuzJtgJRrslrygHKJ3WsZhuCkq+X8c6ivrXDd6zwrLrX3vZP/3PY1yuHHcWR/hEiSlmutpzEQ5XdF+IIz+Uzpeq+gWtMMT1HwIAAA==

Example documentation.

URL

Probably the easiest way to get started to add dashboards to your Grafana instances.

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-from-grafana
spec:
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  url: "https://grafana.com/api/dashboards/1860/revisions/30/download"

NOTE: You don’t have to rely on Grafana Dashboard registry for this, any URL reachable by the operator would work.

Example documentation.

Jsonnet

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-jsonnet
spec:
  resyncPeriod: 30s
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  jsonnet: |
   local grafana = import 'grafonnet/grafana.libsonnet';
   local dashboard = grafana.dashboard;
   local row = grafana.row;
   local singlestat = grafana.singlestat;
   local prometheus = grafana.prometheus;
   local template = grafana.template;

   dashboard.new(
     'JVM',
     tags=['java'],
   )
   .addTemplate(
     grafana.template.datasource(
       'PROMETHEUS_DS',
       'prometheus',
       'Prometheus',
       hide='label',
     )
   )
   .addTemplate(
     template.new(
       'env',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current, env)',
       label='Environment',
       refresh='time',
     )
   )
   .addTemplate(
     template.new(
       'job',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current{env="$env"}, job)',
       label='Job',
       refresh='time',
     )
   )
   .addTemplate(
     template.new(
       'instance',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current{env="$env",job="$job"}, instance)',
       label='Instance',
       refresh='time',
     )
   )
   .addRow(
     row.new()
     .addPanel(
       singlestat.new(
         'uptime',
         format='s',
         datasource='Prometheus',
         span=2,
         valueName='current',
       )
       .addTarget(
         prometheus.target(
           'time() - process_start_time_seconds{env="$env", job="$job", instance="$instance"}',
         )
       )
     )
   )

Plugins

Plugins is a way to extend the grafana functionality in dashboards and datasources.

Plugins can be installed to grafana instances managed by the operator and be defined in both datasources and dashboards.

They cannot be installed using external grafana instances due to how the installation of plugins is done at the start of the instance using environment variables which is a built in feature in grafana.

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: keycloak-dashboard
spec:
  instanceSelector:
    matchLabels:
      dashboards: grafana
  plugins:
    - name: grafana-piechart-panel
      version: 1.3.9
  # NOTE: the json block below is incomplete as it's not the main focus of the example
  json: >
    {
      "__inputs": [
        {
          "name": "DS_PROMETHEUS",
          "label": "Prometheus",
          "description": "",
          "type": "datasource",
          "pluginId": "prometheus",
          "pluginName": "Prometheus"
        }
      ],
    }

Look here for more examples on how to install plugins

Content cache duration

To not constantly perform requests to external URL every time a dashboard reconcile or a resync period expires we save URLs in a cache in the operator. This cache is saved in the status field of the dashboard CR.

By default this cache is 24h long, you can change this value by setting contentCacheDuration manually per dashboard.

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-from-url
spec:
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  url: "https://grafana.com/api/dashboards/7651/revisions/44/download"
  contentCacheDuration: 48h

Remember, depending on where you get your dashboards you might become rate limited if you have multiple dashboards with relatively short contentCacheDuration or if all the requests happens at the same time.

Dashboard uid management

Whenever a dashboard is imported into a Grafana, it gets assigned a random uid unless it’s hardcoded in dashboard’s code. Random uid is undesirable from the operator’s perspective as it would create the need to track those uids across Grafana instances.

To mitigate the scenario, if uid is not hardcoded, the operator will insert the value taken from CR’s metadata.uid (this value is automatically generated by Kubernetes itself for all resources).

Custom folders

In a standard scenario, the operator would use the namespace a CR is deployed to as a folder name in grafana. folder field can be used to set a custom folder name:

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-with-custom-folder
spec:
  folder: "Custom Folder"
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  url: "https://raw.githubusercontent.com/grafana-operator/grafana-operator/master/examples/dashboard_from_url/dashboard.json"

Dashboard customization by providing environment variables

Will be pleasant for scenarios when you would like to extend the behaviour of jsonnet generation by parametrizing it with runtime Env vars:

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: grafanadashboard-jsonnet
spec:
  resyncPeriod: 30s
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  envs:
    - name: API_VERSION
      value: "1.0.0"
    - name: ENV_FROM_CM -- just example, such cm and secrets are not provided by vendor
      valueFrom:
        configMapKeyRef:
          name: custom-grafana-dashboard-cm
          key: GRAFANA_URL
    - name: ENV_FROM_SECRET
      valueFrom:
        secretKeyRef:
          name: custom-grafana-dashboard-secrets
          key: PROMETHEUS_USERNAME
  envFrom: -- just example, such cm and secrets are not provided by vendor
    - configMapRef:
        name: custom-grafana-dashboard-cm
    - secretRef:
        name: custom-grafana-dashboard-secrets
  jsonnet: |
   local grafana = import 'grafonnet/grafana.libsonnet';
   local dashboard = grafana.dashboard;
   local row = grafana.row;
   local singlestat = grafana.singlestat;
   local prometheus = grafana.prometheus;
   local template = grafana.template;
   local labelFromEnvs = std.extVar('API_VERSION');

   dashboard.new(
     'JVM',
     tags=['java'],
   )
   .addTemplate(
     grafana.template.datasource(
       'PROMETHEUS_DS',
       'prometheus',
       'Prometheus',
       hide='label',
     )
   )
   .addTemplate(
     template.new(
       'env',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current, env)',
       label='Environment',
       refresh='time',
     )
   )
   .addTemplate(
     template.new(
       'job',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current{env="$env"}, job)',
       label='Job',
       refresh='time',
     )
   )
   .addTemplate(
     template.new(
       'instance',
       '$PROMETHEUS_DS',
       'label_values(jvm_threads_current{env="$env",job="$job"}, instance)',
       label='Instance',
       refresh='time',
     )
   )
   .addRow(
     row.new()
     .addPanel(
       singlestat.new(
         'uptime',
         format='s',
         datasource='Prometheus',
         span=2,
         valueName='current',
       )
       .addTarget(
         prometheus.target(
           'time() - process_start_time_seconds{env="$env", job="$job", instance="$instance", apiVersion="$labelFromEnvs"}',
         )
       )
     )
   )

Providing runtime to build jsonnet dashboards

This feature provides the ability to pass your jsonnet project with all own or external runtime-required libs/dependencies required in runtime to build your dashboard.

It bridges the signature of the jsonnet generation command jsonnet -J path/to/libs target.jsonnet and Grafana Operator Dashboard CRD. To do this, there are 3 parameters:

jPath - Jsonnet local libs path, must be the same as in your local jsonnet project. Optional part.
fileName - Jsonnet file name which must be built. Required part.
gzipJsonnetProject - Gzip archived project in a byte array representation. Only .tar.gz files are supported. Required part.

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: my-favorite-dashboard-with-internal-dependencies
spec:
  resyncPeriod: 30s
  instanceSelector:
    matchLabels:
      dashboards: "grafana"
  envs:
    - name: API_VERSION
      value: "1.0.0"
    - name: ENV_FROM_CM -- just example, such cm and secrets are not provided by vendor
      valueFrom:
        configMapKeyRef:
          name: custom-grafana-dashboard-cm
          key: GRAFANA_URL
    - name: ENV_FROM_SECRET
      valueFrom:
        secretKeyRef:
          name: custom-grafana-dashboard-secrets
          key: PROMETHEUS_USERNAME
  envFrom: -- just example, such cm and secrets are not provided by vendor
    - configMapRef:
        name: custom-grafana-dashboard-cm
    - secretRef:
        name: custom-grafana-dashboard-secrets
  jsonnetLib:
    jPath:
      - "vendor"
    fileName: "overview.jsonnet"
    gzipJsonnetProject: |-
      {{- (.Files.Get "dashboards.tar.gz") | b64enc | nindent 6 }}

[Example documentation](../examples/dashboard_with_custom_folder/readme).