Configure Grafana datasources and dashboards

A simple role for loading dashboards and datasources for Grafana. It follows the rule that if a dashboard or datasource with the same name already exists, then it is not updated.

It currently works at the organisation level.

Usage

An example of how to use the role is included in example/.

Requires Ansible >=2.2. If your distro doesn't provide this you can run Ansible from a venv:

1. Create venv: $ virtualenv ansible
2. Activate venv: $ source ansible/bin/activate
3. Update pip (optional): $ pip install -U pip
4. Install Ansible: $ pip install ansible

Creating dashboards

To prevent Grafana stripping out datasource names dashboards should be downloaded via the API. This can be done via curl, or via a browser. In this case an example is given for a browser.

In Grafana switch to the organisation in which the dashboard you wish to save is located. Then in the same browser access the dashboard via the API:

http://10.1.2.3:3000/api/dashboards/db/tenant-logs

This will return the dashboard as a JSON string. To allow the dashboard to load successfully it is required to nullify the dashboard id. For example:

<snip>
"dashboard":{
  "annotations":{"list":[]},
  "editable":true",
  "gnetId":null,
  "hideControls":false,
  "id":null,                               <-- Must be null
  "links":[],
  "rows": </snip>