Gatling Timescale DB Publisher

By default, Gatling generates HTML reports for each executed simulation. Individual reports though cannot be compared with previous runs easily. To simplify the comparison of multiple runs and to store results over a longer period of time easy accessible, this repository contains scripts to publish results to a timescale db (PostgreSQL Plugin) to display them in Grafana.

Scripts

• migrate_db.py
  Python script to migrate the database schema to store loadtest results.
  Set environment variable CONNECTION_STRING to connect to the correct database.
  Default is: postgresql://postgres:password@localhost:5432/gatling
  There's no console output if everything finished successfully.
• gatling_timescaledb_publisher.py
  Python script to publish raw data of a simulation.log file to the timescale db.
  Set environment variable CONNECTION_STRING to connect to the correct database.
  Default is: postgresql://postgres:password@localhost:5432/gatling
  Optional script parameter: <simulation.log file name>
• docker/docker-compose.yml
  Docker compose yaml file to start a local timescale db, grafana and pgAdmin. Two dashboards are automatically provisioned to grafana.
  Login credentials:
  • Grafana (port 3000):
    • username: admin
    • password: secret
  • pgAdmin (port 5433):
    • UI
      • pgAdmin-username: test@example.com
      • pgAdmin-password: secret
    • Server-Connection
      • hostname: timescaledb
      • username: postgres
      • password: password

Run locally

The docker-compose.yml file can be started by executing: docker-compose up -d

Install required python dependencies: pip install -r requirements.txt

Initialize the timescale db by running: ./migrate_db.py.

Place the simulation.log file into the directory where the two python scripts are located.

Run the publisher script ./gatling_timescaledb_publisher.py. The script parses the simulation.log file, prints the parsed data, and saves it in the timescale db.

Open a browser and navigate to http://localhost:3000. Login with the grafana credentials as described above.

The docker containers can be stopped and removed by running: docker-compose down

Run remotely

Set CONNECTION_STRING environment variable with a valid postgresql connection string for the timescale db.

Execute the scripts as described above.

Python versions

The script was developed and tested with Python 3.8.

License

The code in this repository is licensed under the MIT License. The used libraries under their respective licenses (see requirements.txt).