pytest-resource-usage
Add running times and peak memory usage of tests to the output of
pytest. Two ways of measuring memory usage are
implemented. The first uses tracemalloc, which is unaffected by
swapping. Since tracemalloc is a standard library, this way of
measuring introduces no dependencies beyond pytest. However, measuring
with tracemalloc can have a high overhead. The other way of measuring
memory usage is by means of sampling the unique set size. This method
has lower overhead, but can be less accurate and is only available if
the psutil package is installed.
The pytest-resource-usage package is intentionally kept fairly simple.
If you want something more feature-rich and ambitious, you should use
pytest-monitor.
Example usage
Reporting is triggered by the presence of one or more of the following markers:
-
report_duration, -
report_tracemalloc, -
report_uss(requirespsutil).
from time import sleep
import pytest
@pytest.mark.report_duration
def test_sleep():
sleep(99)
@pytest.mark.report_uss
@pytest.mark.report_tracemalloc
@pytest.mark.report_duration
@pytest.mark.parametrize("elements", [2_000_000, 1_000_000])
def test_tracemalloc_overhead(elements):
_ = list(range(elements))
@pytest.mark.report_uss(interval=0.01)
@pytest.mark.report_duration
def test_unique_set_size():
_ = list(range(1_000_000))
@pytest.mark.report_duration
def test_no_overhead():
_ = list(range(1_000_000))Running the above tests produces the following pytest output.
================================================================ test session starts =================================================================
plugins: pytest-resource-usage-1.0.0
collected 5 items
tests/test_readme.py ..... [100%]
=================================================================== resource usage ===================================================================
tests/test_readme.py::test_sleep (call) running time: 0:01:39
tests/test_readme.py::test_tracemalloc_overhead[2000000] (call) peak allocated memory: 72MB, peak unique set size: 243MB, running time: 0.627 seconds
tests/test_readme.py::test_tracemalloc_overhead[1000000] (call) peak allocated memory: 36MB, peak unique set size: 115MB, running time: 0.329 seconds
tests/test_readme.py::test_unique_set_size (call) peak unique set size: 25.7MB, running time: 0.047 seconds
tests/test_readme.py::test_no_overhead (call) running time: 0.020 seconds
=========================================================== 5 passed in 100.13s (0:01:40) ============================================================
