How To Use

Create a test file in the tests folder

There are three ways to execute the tests:

1. Running tests with an external Splunk instance

pip3 install pytest-splunk-addon

Run pytest with the add-on, in an external Splunk deployment

pytest --splunk-type=external --splunk-app=<path-to-addon-package> --splunk-data-generator=<path to pytest-splunk-addon-data.conf file> --splunk-host=<hostname> --splunk-port=<splunk-management-port> --splunk-user=<username> --splunk-password=<password> --splunk-hec-token=<splunk_hec_token>

2. Running tests with docker splunk

git clone git@github.com:splunk/pytest-splunk-addon.git
cd pytest-splunk-addon
pip install poetry
poetry install

Create a Dockerfile.splunk file

Create a Dockerfile.uf file

Create docker-compose.yml

Create conftest.py in the test folder along with the test file

Run pytest with the add-on, using the following command:

pytest --splunk-type=docker --splunk-data-generator=<path to pytest-splunk-addon-data.conf file>

The tool assumes the Splunk Add-on is located in a folder “package” in the project root.

Note

  • If live events are available in external Splunk instance or docker splunk, then SA-Eventgen is not required. This is applicable only till v1.2.0 of pytest-splunk-addon.

  • From v1.3.0 pytest-splunk-addon ingests data independently which is used for execution of all the test cases.

3. Running tests with an external forwarder and Splunk instance

  • Run pytest with the add-on, using an external forwarder sending events to another Splunk deployment where a user can search for received events.

  • Forwarding & receiving configuration in –splunk-forwarder-host and –splunk-host must be done before executing the tests.

  • User can validate the forwarding using makeresults command.

| makeresults | eval _raw="sample event" | collect index=main, source=test_source, sourcetype=test_src_type
  • Sample pytest command with the required params

pytest --splunk-type=external                                   # Whether you want to run the addon with docker or an external Splunk instance
    --splunk-app=<path-to-addon-package>                        # Path to Splunk app package. The package should have the configuration files in the default folder.
    --splunk-host=<hostname>                                    # Receiver Splunk instance where events are searchable.
    --splunk-port=<splunk_management_port>                      # default 8089
    --splunk-user=<username>                                    # default admin
    --splunk-password=<password>                                # default Chang3d!
    --splunk-forwarder-host=<splunk_forwarder_host>             # Splunk instance where forwarding to receiver instance is configured.
    --splunk-hec-port=<splunk_forwarder_hec_port>               # HEC port of the forwarder instance.
    --splunk-hec-token=<splunk_forwarder_hec_token>             # HEC token configured in forwarder instance.
    --splunk-data-generator=<pytest_splunk_addon_conf_path>     # Path to pytest-splunk-addon-data.conf

Note

  • Forwarder params are supported only for external splunk-type.

  • If Forwarder params are not provided It will ingest and search in the same Splunk deployment provided in –splunk-host param.


There are 3 types of tests included in pytest-splunk-addon are:

  1. To generate test cases only for knowledge objects, append the following marker to pytest command:

    -m  splunk_searchtime_fields
    
  2. To generate test cases only for CIM compatibility, append the following marker to pytest command:

    -m  splunk_searchtime_cim
    
  3. To generate test cases only for index time properties, append the following marker to pytest command:

    -m  splunk_indextime --splunk-data-generator=<Path to the conf file>
    

    For detailed information on index time test execution, please refer here.

  • To execute all the searchtime tests together, i.e both Knowledge objects and CIM compatibility tests, append the following marker to the pytest command:

    -m  "splunk_searchtime_fields or splunk_searchtime_cim"
    

The following optional arguments are available to modify the default settings in the test cases:

  1. To search for events in a specific index, user can provide following additional arguments:

    --search-index=<index>
    
        Splunk index of which the events will be searched while testing. Default value: "*, _internal".
    
  2. To increase/decrease time interval and retries for flaky tests, user can provide following additional arguments:

    --search-retry=<retry>
    
        Number of retries to make if there are no events found while searching in the Splunk instance. Default value: 0.
    
    --search-interval=<interval>
    
        Time interval to wait before retrying the search query.Default value: 0.
    
  3. To discard the eventlog generation in the working directory, user can provide following additional argument along with pytest command:

    --discard-eventlogs
    
  4. To enable the Splunk Index cleanup performed before the test run, user can provide argument along with pytest command:

    --splunk-cleanup
    
  5. A new functionality is introduced in pytest-splunk-addon to suppress unwanted errors in test_splunk_internal_errors.

    • Splunk related errors: There is a file maintained in pytest-splunk-addon “.ignore_splunk_internal_errors” , user can add the string in the file and events containing these strings will be suppressed by the search query.

    • Addon related errors: To suppress these user can create a file with the list of strings and provide the file in the –ignore-addon-errors param while test execution.

    --ignore-addon-errors=<path_to_file>
    
    • Sample strings in the file.

    SearchMessages - orig_component="SearchStatusEnforcer"
    message_key="" message=NOT requires an argument
    

    Note

    Each line in the file will be considered a separate string to be ignored in the events.

    • Sample Event which will be ignored by the search query.

    11-04-2020 13:26:01.026 +0000 ERROR SearchMessages - orig_component="SearchStatusEnforcer" app="search" sid="ta_1604496283.232" peer_name="" message_key="" message=NOT requires an argument
    

Extending pytest-splunk-addon

1. Test cases taking too long to execute

Use pytest-xdist to execute test cases across multiple processes.

How to use pytest-xdist :

  • pip install pytest-xdist

  • add -n {number-of-processes} to the pytest command

This will create the mentioned amount of processes and divide the test cases amongst them.

Note

Make sure there is enough data on the Splunk instance before running tests with pytest-xdist because faster the execution, lesser the time to generate enough data.

2. Want flaky/known failures to not fail the execution

Use pytest-expect to mark a list of test cases as flaky/known failures which will not affect the final result of testing.

How to use pytest-expect:

  • pip install pytest-expect

  • Add --update-xfail to the pytest command to generate a .pytest.expect file, which is a list of failures while execution.

  • Make sure that the .pytest.expect file is in the root directory from where the test cases are executed.

  • When the test cases are executed the next time, all the tests in the .pytest.expect file will be marked as xfail 1

  • If there is a custom file containing the list of failed test cases, it can be used by adding --xfail-file custom_file to the pytest command.

Note

Test cases should be added to .pytest.expect only after proper validation.

3. Setup test environment before executing the test cases

If any setup is required in the Splunk/test environment before executing the test cases, implement a fixture in conftest.py.

@pytest.fixture(scope="session")
def splunk_setup(splunk):
    # Will be executed before test execution starts
    . . .

The setup fixture opens many possibilities to setup the testing environment / to configure Splunk. For example,

  • Enable Saved-searches

  • Configure the inputs of an Add-on.

  • Wait for an lookup to be populated.

  • Restart Splunk.

The following snippet shows an example in which the setup fixture is used to enable a saved search.

4. Check mapping of an add-on with custom data models

pytest-splunk-addon is capable of testing mapping of an add-on with custom data models.

How can this be achieved :

  • Make json representation of the data models, which satisfies this DataModelSchema.

  • Provide the path to the directory having all the data models by adding --splunk_dm_path path_to_dir to the pytest command

  • The test cases will now be generated for the data models provided to the plugin and not for the default data models.


1

xfail indicates that you expect a test to fail for some reason. A common example is a test for a feature not yet implemented, or a bug not yet fixed. When a test passes despite being expected to fail, it’s an xpass and will be reported in the test summary.