Pytest-monitor is a pytest plugin designed for analyzing resource usage.
- Analyze your resources consumption through test functions:
- memory consumption
- time duration
- CPU usage
- Keep a history of your resource consumption measurements.
- Compare how your code behaves between different environments.
Simply run pytest as usual: pytest-monitor is active by default as soon as it is installed. After running your first session, a .pymon sqlite database will be accessible in the directory where pytest was run.
Example of information collected for the execution context:
ENV_H | CPU_COUNT | CPU_FREQUENCY_MHZ | CPU_TYPE | CPU_VENDOR | RAM_TOTAL_MB | MACHINE_NODE | MACHINE_TYPE | MACHINE_ARCH | SYSTEM_INFO | PYTHON_INFO |
---|---|---|---|---|---|---|---|---|---|---|
8294b1326007d9f4c8a1680f9590c23d | 36 | 3000 | x86_64 | Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz | 772249 | some.host.vm.fr | x86_64 | 64bit | Linux - 3.10.0-693.el7.x86_64 | 3.6.8 (default, Jun 28 2019, 11:09:04) n[GCC ... |
Here is an example of collected data stored in the result database:
RUN_DATE | ENV_H | SCM_ID | ITEM_START_TIME | ITEM | KIND | COMPONENT | TOTAL_TIME | USER_TIME | KERNEL_TIME | CPU_USAGE | MEM_USAGE |
---|---|---|---|---|---|---|---|---|---|---|---|
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:36.890477 | pkg1.test_mod1/test_sleep1 | function | None | 1.005669 | 0.54 | 0.06 | 0.596618 | 1.781250 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:39.912029 | pkg1.test_mod1/test_heavy[10-10] | function | None | 0.029627 | 0.55 | 0.08 | 21.264498 | 1.781250 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:39.948922 | pkg1.test_mod1/test_heavy[100-100] | function | None | 0.028262 | 0.56 | 0.09 | 22.998773 | 1.781250 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:39.983869 | pkg1.test_mod1/test_heavy[1000-1000] | function | None | 0.030131 | 0.56 | 0.10 | 21.904277 | 2.132812 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:40.020823 | pkg1.test_mod1/test_heavy[10000-10000] | function | None | 0.060060 | 0.57 | 0.14 | 11.821601 | 41.292969 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:40.093490 | pkg1.test_mod2/test_sleep_400ms | function | None | 0.404860 | 0.58 | 0.15 | 1.803093 | 2.320312 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:40.510525 | pkg2.test_mod_a/test_master_sleep | function | None | 5.006039 | 5.57 | 0.15 | 1.142620 | 2.320312 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:45.530780 | pkg3.test_mod_cl/test_method1 | function | None | 0.030505 | 5.58 | 0.16 | 188.164762 | 2.320312 |
2020-02-17T09:11:36.731233 | 8294b1326007d9f4c8a1680f9590c23d | de23e6bdb987ae21e84e6c7c0357488ee66f2639 | 2020-02-17T09:11:50.582954 | pkg4.test_mod_a/test_force_monitor | function | test | 1.005015 | 11.57 | 0.17 | 11.681416 | 2.320312 |
A full documentation is available.
You can install pytest-monitor via conda (through the conda-forge channel):
$ conda install pytest-monitor -c https://conda.anaconda.org/conda-forge
Another possibility is to install pytest-monitor via pip from PyPI:
$ pip install pytest-monitor
You will need a valid Python 3.5+ interpreter. To get measures, we rely on:
- psutil to extract CPU usage
- memory_profiler to collect memory usage
- and pytest (obviously!)
Note: this plugin doesn't work with unittest
By default, pytest-monitor stores its result in a local SQLite3 local database, making results accessible. If you need a more powerful way to analyze your results, checkout the monitor-server-api which brings both a REST Api for storing and historize your results and an API to query your data. An alternative service (using MongoDB) can be used thanks to a contribution from @dremdem: pytest-monitor-backend.
Contributions are very welcome. Tests can be run with tox. Before submitting a pull request, please ensure that:
- both internal tests and examples are passing.
- internal tests have been written if necessary.
- if your contribution provides a new feature, make sure to provide an example and update the documentation accordingly.
This code is distributed under the MIT license. pytest-monitor is free, open-source software.
If you encounter any problem, please file an issue along with a detailed description.
The main author of pytest-monitor is Jean-SĂ©bastien Dieu, who can be reached at jean-sebastien.dieu@cfm.fr.
This pytest plugin was generated with Cookiecutter along with @hackebrot's cookiecutter-pytest-plugin template.