Log Metrics

The log_metric fixture is a powerful tool for recording performance metrics or other numerical data during your tests. It generates a file that follows the Prometheus text-based format, which is highly compatible with modern monitoring systems and the OpenMetrics standard.

Use Case

You can use this fixture to track key metrics from your embedded device, such as boot time, memory usage, or network throughput. By logging these values, you can monitor performance trends over time and catch regressions automatically.

CLI Argument

To enable metric logging, you need to provide the --metric-path command-line argument. This specifies the file where the metrics will be saved.

pytest --metric-path=output/metrics.txt

Fixture Usage

To use the fixture, simply include log_metric as an argument in your test function. It provides a callable that you can use to log your metrics.

def test_my_app(log_metric):
    # ... test code ...
    boot_time = 123.45  # measured boot time
    log_metric("boot_time", boot_time, target="esp32", sdk="v5.1")

Output Format

The metrics are written to the file specified by --metric-path in the Prometheus text-based format. Each line represents a single metric.

Example output in output/metrics.txt:

boot_time{target="esp32",sdk="v5.1"} 123.45

If --metric-path is not provided, the log_metric function will do nothing and issue a UserWarning.