Druid as a Data Source for Grafana

Table of Contents

• What is this Druid-Grafana plugin ?
• Features
• Screenshots
  • Datasource - connection
  • Datasource - default settings
  • Panels
  • Query builder - JSON
  • Query builder - SQL
  • Query builder - timeseries
  • Query builder - settings
  • Variables
  • Variables - formatter - druid:json
  • Explore - logs
  • Query builder - settings - logs
• Contribute
• Install
• Examples

What is this Druid-Grafana plugin?

Grafana doesn't supports Druid as a Data Source, this plugin aims to do so.

Features

At the time of writing, the plugin supports all Grafana features and all Druid queries:

• Druid queries: SQL, timeseries, topn, groupby, timeboundary, segmentmetadata, datasourcemetadata, scan, search, JSON
• Variables: Grafana global variables replacement, query variables, formatter druid:json (provide support for multi-value variables within rune queries).
• Alerts
• Explore
• Logs

if you're using a self-signed TLS certificate, an option to "Skip TLS verify" will be shown when "https" is used in datasource URI)

Screenshots

Datasource - Connection

Datasource - Default settings

Panels

Query builder - JSON

Query builder - SQL

Query builder - Timeseries

Query builder - Settings

Variables

Variables - Formatter - druid:json

Explore - Logs

Query builder - Settings - Logs

Contribute

Any contribution is welcome! Feel free to join on Slack to discuss :) To go further, see Contribute

Install

• From Grafana.com: grafana-cli plugins install grafadruid-druid-datasource $VERSION
  • If $VERSION is not specified latest version will be installed
• From remote archive: grafana-cli --pluginUrl https://github.com/grafadruid/druid-grafana/releases/download/v$VERSION/grafadruid-druid-datasource-$VERSION.zip plugins install grafadruid-druid-datasource
• From local archive: unzip grafadruid-druid-datasource-$VERSION.zip -d $YOUR_PLUGIN_DIR

Where $VERSION is for instance 1.0.0 and $YOUR_PLUGIN_DIR is for instance /var/lib/grafana/plugins

(Source: https://grafana.com/docs/grafana/latest/plugins/installation/)

Build lolo-druid plugin on Mac and Install it via Kubernetes manually

Build lolo-druid plugin

Tested on Mac Pro M1 - Monterey v12.6:

• Launch Desktop Docker
• Build

./mage-macos buildAll

Copy the plugin into a Grafana Pod in Kubernetes

• Make sure you have set a proper Kubernetes context
• Rename the dist folder into like lolo-grafadruid-druid-datasource
• Copy the plugin into a Grafana Pod in Kubernetes (e.g.,)

$ kubectl get pod -n <yourNamespace>
NAME                       READY   STATUS    RESTARTS   AGE
grafana-7c549965bc-47trw   1/1     Running   0          1d

$ kubectl cp -n <yourNamespace> ./dist grafana-7c549965bc-47trw:/var/lib/grafana/plugins/lolo-grafadruid-druid-datasource

• Make sure your Grafana pod has the env variable: GF_PLUGINS_ALLOW_LOADING_UNSIGNED_PLUGINS : lolo-grafadruid-druid-datasource
• Restart the pod and check if lolo-druid is installed :)

Examples

You can try out various advanced features of the plugin by importing the demo dashboard and running it against the Wikipedia dataset used in the Druid quickstart tutorial.

When using the provided docker based environment (./mage env:start && ./mage), the dataset is automatically ingested in Druid, the datasource and dashboard are automatically provisionned in Grafana.