Executing Performance Tests¶

This repository includes tools to execute performance tests for the APEx Dispatch API. Performance testing is useful for analyzing the impact of code changes, database updates, and platform modifications on the system's behavior and responsiveness.

Prerequisites¶

Before running the performance tests, ensure the following prerequisites are met:

Python environment (Python 3.10+ recommended)
Docker and Docker Compose installed on your system

Setting Up the Environment¶

Performance tests require both the API and a database to be running locally. Follow these steps to set up your environment:

Create a .env file in the root of the repository with the following variables:
OPENEO_BACKENDS_PERFORMANCE → Configuration for the openEO backend authentication. See configuration guide.
KEYCLOAK_CLIENT_PERFORMANCE_ID → Client ID used for executing performance tests.
KEYCLOAK_CLIENT_PERFORMANCE_SECRET → Client secret used for executing performance tests.
Start the services using Docker Compose:

docker compose -f docker-compose.perf.yml up -d db

Starts a local database instance.

docker compose -f docker-compose.perf.yml up -d migrate

Executes database migrations to ensure all required tables are created.

docker compose -f docker-compose.perf.yml up -d app

Starts the API locally.

Tip: You can check the logs of each service with docker compose -f docker-compose.perf.yml logs -f <service_name>.

Executing Performance Tests¶

The performance tests are implemented using Locust. Test scenarios are located in tests/performance/locustfile.py.

Running Tests with a Web Dashboard¶

To execute the performance tests and monitor them in a browser dashboard:

locust -f tests/performance/locustfile.py -u 10 --host http://localhost:8000 --run-time 1m

-u 10 → Number of simulated concurrent users
--host http://localhost:8000 → URL of the API to test
--run-time 1m → Duration of the test

After starting, open your browser at http://localhost:8089 to monitor real-time performance metrics, including response times, failure rates, and throughput.

Running Tests in Headless Mode¶

To execute tests without a web interface (useful for CI/CD pipelines):

locust -f tests/performance/locustfile.py -u 10 --host http://localhost:8000 --run-time 1m --headless

You can also export the results to a CSV file for further analysis:

locust -f tests/performance/locustfile.py -u 10 --host http://localhost:8000 --run-time 1m --headless --csv=perf_test_results

Recommended Practices¶

Start with a small number of users to validate test scripts before scaling up.
Combine performance testing with monitoring tools to detect resource bottlenecks.