bureau14 / qdb-grafana-http-adapter

Quasardb Grafana HTTP Adapter

Requirements

• Windows, Linux, Freebsd, macOS
• quasardb Daemon running (qdbd)
• qdb-api-python installed (https://github.com/bureau14/qdb-api-python)
• Grafana Server running
• SimpleJSON (for Grafana) plugin installed

How to use it ?

Getting started

By default only timeseries with tag grafana will be listed in your grafana interface. You can edit this tag, replacing settings.py/[TAG_TS_GRAFANA].

Update QDB_CLUSTER_URI variable in settings.py w/ URI of your Quasardb cluster. Default PORT_API is 8080, can be replaced in settings.py.

Start the API

pip install -r requirements.txt

python app.py

In Grafana

Accessing grafana interface

Go to grafana web interface (default is http://localhost:3000) Default credentials (admin/admin)

Then you should see an interface similar to the below

Requirements

SimpleJSON is required to crunch the data from the API

Go to plugins section, then Find more plugins on Grafana.com, SimpleJSON should be in the list of plugins

Adding a Data Source

Once SimpleJSON plugin is installed, go to Data sources section, then click on the Add data source link

Under config tab :

• select a name for your data source
• select SimpleJson as a type
• use the URI of the API as a Data Source : http://$(MY_HOST):$(PORT_API)
• then click on Save & Test

Displaying the Data

• click on Create a new Dashboard under home view panel
• then, click on the pictogram graph or table
• click on Widget label to access configuration (e.g : 'Panel Title')
• then on Edit button
• You'll see a dropdown to select your data source, and another one to add one or more Timeserie target(s)

Warning

If you use a special character such as :, ,, % in your timeseries name nor columns name, you will need change the delimiter settings.py/[SEPARATOR_TS_LABEL]; it will be the pattern to keep out of columns/timeseries aliases.

To improve performances on grafana panel, we've set a variable in settings.py/[ARBITRATY_NB_DATAPOINTS], which'll limit the number of datapoints in API Output. If you don't want to limit the output, set settings.py/[ARBITRATY_NB_DATAPOINTS] to None.

Note for macOS users

qdb-api-python is not yet totally supported on macOS X, so consider running the API in a supported environement such as Windows, Linux or FreeBSD.