Customizable dashboards

Introduced in GitLab 15.5 as an experiment.

Customizable dashboards provide a configuation-based dashboard structure, which is used to render and modify dashboard configurations created by GitLab or users.

The dashboard structure does not provide the means to save and version control user configuration files on a repository. This functionality is provided by Analytics dashboards which uses the customizable dashboard component.

NOTE: Customizable dashboards is intended for Premium and Ultimate subscriptions.

Overview

Customizable dashboard can be broken down into 3 logical components:

Dashboard
Panels
Visualizations

A dashboard consists of a list of panels to render, each panel references one visualization, and a single visualization can be shared by many panels.

A typical dashboard structure looks like this:

dashboard
├── panelA
│  └── visualizationX
├── panelB
│  └── visualizationY
├── panelC
│  └── visualizationY

Usage

To use customizable dashboards:

Create a new Vue component for your dashboard.
Create a visualization configuration.
Create your dashboard configuration.
Render an instance of CustomizableDashboard and pass it your dashboard configuration.

Visualization configuration

Each visualization is a graphical representation of query results fetched from a data source.

// visualizations.js

export const pageViewsOverTime = {
  // The name of the Vue component used to render the query.
  type: 'LineChart',
  // Chart options defined by the charting library being used by the panel.
  options: {
    xAxis: { name: __('Time'), type: 'time' },
    yAxis: { name: __('Counts'), type: 'value' },
  },
  // The data to query
  data: {
    // The data source to query. Here it is Product Analytics.
    type: 'cube_analytics',
    // The query to run on the data source. Here in Cube.js format.
    query: {
      dimensions: [],
      filters: [
        {
          member: 'TrackedEvents.eventName',
          operator: 'equals',
          values: ['page_view']
        }
      ],
      measures: ['TrackedEvents.pageViewsCount'],
      timeDimensions: [
        {
          dimension: 'TrackedEvents.derivedTstamp',
          granularity: 'day',
        },
      ],
      limit: 100,
      timezone: 'UTC',
    },
  },
};

Adding a new visualization render type

To add a new visualization render type:

Create a new Vue component that accepts data and options properties. See line_chart.vue as an example.
Add your component to the list of conditional imports in panel_base.vue.
Add your component to the schema's list of AnalyticsVisualization types in analytics_visualizations.json.

Migrating existing components to visualizations

You can migrate existing components to dashboard visualizations. To do this, wrap your existing component in a new visualization that provides the component with the required context and data. See dora_performers_score.vue as an example.

As an upgrade path, your component may fetch its own data internally. This method is fine for the first few iterations, but eventually you should migrate your visualization to use the standard and shared analytics data sources method. See value_stream.js as an example.

Adding a new visualization data source

To add a new data source:

Create a new JavaScript module that exports a fetch method. See cube_analytics.js as an example.
Add your module to the list exports in data_sources/index.js.
Add your data source to the schema's list of Data types in analytics_visualizations.json.

NOTE: Your data source must respect the filters so that all panels show data for the same date range.

Dashboard configuration

Here is an example dashboard configuration:

// constants.js
import { pageViewsOverTime } from './visualizations';

export const dashboard = {
  slug: 'my_dashboard', // Used to set the URL path for the dashboard.
  title: 'My dashboard title', // The title to display.
  description: 'This is a description of the dashboard', // A description of the dashboard
  userDefined: true, // The dashboard editor is only available when true.
  // Each dashboard consists of an array of panels to display.
  panels: [
    {
      id: 1,
      title: 'Page views over time', // The panel title to display.
      // The visualization configuration. This can be shared by many panels.
      visualization: pageViewsOverTime,
      // Gridstack settings based upon https://github.com/gridstack/gridstack.js/tree/master/doc#item-options.
      // All values are grid row/column numbers up to 12.
      // We use the default 12 column grid https://github.com/gridstack/gridstack.js#change-grid-columns.
      gridAttributes: {
        yPos: 1,
        xPos: 0,
        width: 6,
        height: 5,
      },
      // Optional overrides for the values in `visualization.data.query`.
      // Here we override the Cube.js query to get page views per week instead of days.
      queryOverrides: {
        timeDimensions: {
          dimension: 'TrackedEvents.derivedTstamp',
          granularity: 'week',
        },
      },
    },
  ],
};

Using the component

Here is an example component that renders a customizable dashboard:

<script>
import CustomizableDashboard from 'ee/vue_shared/components/customizable_dashboard/customizable_dashboard.vue';
import PanelsBase from `ee/vue_shared/components/customizable_dashboard/panels_base.vue`;
import { dashboard } from './constants';

export default {
  name: 'MyCustomDashboard',
  components: {
    CustomizableDashboard,
    PanelsBase,
  },
  data() {
    return {
      // We keep the original (default) dashboard object to track changes.
      dashboard: {
        ...dashboard,
        default: { ...dashboard, }
      },
      // Optional dashboard filters. Currently the only availble filter is date range.
      defaultFilters: {
        dateRangeOption: 'last_7_days' // 'custom', 'today', 'last_7_days', 'last_30_days'
        endDate: new Date(2023, 06, 14),
        startDate: new Date(2023, 06, 7),
      },
      // Set to true to sync the filter object with the URL query string.
      syncUrlFilters: true,
      // Set to true to show the date range filter.
      showDateRangeFilter: true,
      // The maximum size of the date range allowed in days. 0 for unlimited.
      dateRangeLimit: 0,
      // Array of GlDisclosureDropdown items to show on each panel when editing
      panelActions: [
        {
          text: __('Delete'),
          action: () => this.$emit('delete'),
          icon: 'remove',
        },
      ],
    };
  },
};
</script>

<template>
  <customizable-dashboard
    :initial-dashboard="dashboard"
    :default-filters="defaultFilters"
    :sync-url-filters="syncUrlFilters"
    :show-date-range-filter="showDateRangeFilter"
    :date-range-limit="dateRangeLimit"
  />
    <template #panel="{ panel, filters, editing, deletePanel }">
      <!-- Panels base provides a styled wrapper for your visualizations. -->
      <panels-base
        :title="panel.title"
        :editing="editing"
        :actions="panelActions"
        @delete="deletePanel"
      >
        <template #body>
          <!-- Render the panel's visualization here -->
        </template>
      </panels-base>
    </template>
  </customizable-dashboard>
</template>

Dashboard designer

Introduced in GitLab 16.1 with a flag named combined_analytics_dashboards_editor. Disabled by default.

Generally available in GitLab 16.6. Feature flag combined_analytics_dashboards_editor removed.

The CustomizableDashboard component provides a graphical interface for users to modify panels of existing dashboards and create new dashboards.

The dashboard editor is only available when dashboard.userDefined is true.

<script>
import CustomizableDashboard from 'ee/vue_shared/components/customizable_dashboard/customizable_dashboard.vue';
import { s__ } from '~/locale';
import { dashboard } from './constants';

export const I18N_MY_NEW_CATEGORY = s__('Namespace|My data source');

export default {
  name: 'MyCustomDashboard',
  data() {
    return {
      ...,
      // The initial saved dashboard. Used to track changes.
      initialDashboard: dashboard,
      // Set to true to render the dashboard saving state.
      isSaving: false,
      // A list of availble visualizations categorized by feature.
      availableVisualizations: {
        // The visualization category title to display.
        [I18N_MY_NEW_CATEGORY]: {
          // Set to true when asynchronously loading the visualization IDs
          loading: false,
          // List of availble visualization IDs for the user to add to the dashboard.
          visualizationIds: [
            'page_views_over_time',
            'events_over_time',
          ],
        },
      }
    };
  },
  methods: {
    /**
     * Event handler for when a user saves changes made to the current dashboard.
     * @param  {String} dashboardId The current dashboard ID.
     * @param  {String} newDashboardObject The newly modified dashboard object.
     */
    saveDashboard(dashboardId, newDashboardObject) {
      this.isSaving = true;
      // Save changes somewhere.
      // Then update the saved dashboard version
      this.initialDashboard = newDashboardObject;
      this.isSaving = false;
    },
  },
}
</script>

<template>
  <customizable-dashboard
    :initial-dashboard="initialDashboard"
    :available-visualizations="availableVisualizations"
    :is-saving="isSaving"
    @save="saveDashboard"
  />
    <template #panel="{ panel, filters, editing, deletePanel }">
      <my-dashboard-panel :panel="panel" />
    </template>
  </customizable-dashboard>
</template>

Introducing visualizations behind a feature flag

While developing new visualizations we can use feature flags to mitigate risks of disruptions or incorrect data for users.

The from_data method builds the panel objects for a dashboard. Using the filter_map method, we can add a condition to skip rendering panels that include the visualization we are developing.

For example, here we have added the enable_usage_overview_visualization feature flag and can check it's current state to determine whether panels using the usage_overview visualization should be rendered:

panel_yaml.filter_map do |panel|
  # Skip processing the usage_overview panel if the feature flag is disabled
  next if panel['visualization'] == 'usage_overview' && Feature.disabled?(:enable_usage_overview_visualization)

  new(
    title: panel['title'],
    project: project,
    grid_attributes: panel['gridAttributes'],
    query_overrides: panel['queryOverrides'],
    visualization: panel['visualization']
  )
end