A plugin to send pytest test results to ELK stack, with extra context data

• Report each test result into Elasticsearch as they finish
• Automatically append contextual data to each test:
  • git information such as branch or last commit and more
  • all of CI env variables
    • Jenkins
    • Travis
    • Circle CI
    • Github Actions
  • username if available
• Report a test summary to Elastic for each session with all the context data
• Append any user data into the context sent to Elastic

• having pytest tests written

You can install "pytest-elk-reporter" via pip from PyPI

pip install pytest-elk-reporter

We need this auto_create_index setting enabled for the indexes that are going to be used, since we don't have code to create the indexes, this is the default

curl -X PUT "localhost:9200/_cluster/settings" -H 'Content-Type: application/json' -d'
{
    "persistent": {
        "action.auto_create_index": "true"
    }
}
'

For more info on this Elasticsearch feature check their index documention

pytest --es-address 127.0.0.1:9200
# or if you need user/password to authenticate
pytest --es-address my-elk-server.io:9200 --es-username fruch --es-password 'passwordsarenicetohave'

from pytest_elk_reporter import ElkReporter

def pytest_plugin_registered(plugin, manager):
    if isinstance(plugin, ElkReporter):
      # TODO: get credentials in more secure fashion programmatically, maybe AWS secrets or the likes
      # or put them in plain-text in the code... what can ever go wrong...
      plugin.es_address = "my-elk-server.io:9200"
      plugin.es_user = 'fruch'
      plugin.es_password = 'passwordsarenicetohave'
      plugin.es_index_name = 'test_data'

# put this in pytest.ini / tox.ini / setup.cfg
[pytest]
es_address = my-elk-server.io:9200
es_user = fruch
es_password = passwordsarenicetohave
es_index_name = test_data

see pytest docs for more about how to configure pytest using .ini files

In this example, I'll be able to build a dashboard for each version:

import pytest

@pytest.fixture(scope="session", autouse=True)
def report_formal_version_to_elk(request):
    """
    Append my own data specific, for example which of the code under test is used
    """
    # TODO: programmatically set to the version of the code under test...
    my_data = {"formal_version": "1.0.0-rc2" }

    elk = request.config.pluginmanager.get_plugin("elk-reporter-runtime")
    elk.session_data.update(**my_data)

import requests

def test_my_service_and_collect_timings(request, elk_reporter):
    response = requests.get("http://my-server.io/api/do_something")
    assert response.status_code == 200

    elk_reporter.append_test_data(request, {"do_something_response_time": response.elapsed.total_seconds() })
    # now, a dashboard showing response time by version should be quite easy
    # and yeah, it's not exactly a real usable metric, but it's just one example...

Or via the record_property built-in fixture (that is normally used to collect data into junit.xml reports):

import requests

def test_my_service_and_collect_timings(record_property):
    response = requests.get("http://my-server.io/api/do_something")
    assert response.status_code == 200

    record_property("do_something_response_time", response.elapsed.total_seconds())

One cool thing that can be done now that you have a history of the tests, is to split the tests based on their actual runtime when passing. For long-running integration tests, this is priceless.

In this example, we're going to split the run into a maximum of 4 min slices. Any test that doesn't have history information is assumed to be 60 sec long.

# pytest --collect-only --es-splice --es-max-splice-time=4 --es-default-test-time=60
...

0: 0:04:00 - 3 - ['test_history_slices.py::test_should_pass_1', 'test_history_slices.py::test_should_pass_2', 'test_history_slices.py::test_should_pass_3']
1: 0:04:00 - 2 - ['test_history_slices.py::test_with_history_data', 'test_history_slices.py::test_that_failed']

...

# cat include000.txt
test_history_slices.py::test_should_pass_1
test_history_slices.py::test_should_pass_2
test_history_slices.py::test_should_pass_3

# cat include000.txt
test_history_slices.py::test_with_history_data
test_history_slices.py::test_that_failed

### now we can run each slice on its own machine
### on machine1
# pytest $(cat include000.txt)

### on machine2
# pytest $(cat include001.txt)

Contributions are very welcome. Tests can be run with tox. Please ensure the coverage at least stays the same before you submit a pull request.

Distributed under the terms of the MIT license, "pytest-elk-reporter" is free and open source software

If you encounter any problems, please file an issue along with a detailed description.

This pytest plugin was generated with Cookiecutter along with @hackebrot's cookiecutter-pytest-plugin template.