NGI Dashboards

This code is for use in monitoring key performance indicators (KPIs) at the SciLifeLab National Genomics Infrastructure in Stockholm, Sweden.

See the documentation for a little more information.

Method Outline

1. A script polls statusdb for a range of KPI statistics. These are logged to the innovatively titled kpi database within statusdb.
   • The script runs on a cronjob, once per hour.
   • It can be run from anywhere with access to statusdb, but this will typically be on the tools server.
   • Genomics Status can also access this data for plotting.
2. A second script pulls information from the kpi database and uses it to parse some templates using the jinja2 package.
   • At the time of writing, there are three templates - one for the internal dashboard, one for the external dashboard, and a HTML chunk to be displayed on the external website.
   • This script will also be run on a cron job, scheduled to run a few minutes after the one described above.
3. The dashboards work with raspberry PIs. These have a copy of Firefox running which will poll the apache server on tools. This will serve up the static HTML file generated in step 2.

Updating KPI Targets

The KPI targets are stored in a config file held in this repository. To change them, fork the repository and change the file. Submit a pull request and merge the changes. Then pull these changes to tools.

Installation

conda create -n dashboard python=2 matplotlib
conda activate dashboard
pip install -r requirements.txt

Contact

This code was developed by Phil Ewels (@ewels) and Remi-Andre Olsen (@remiolsen). Feel free to get in touch if you have any queries.