Marconi's system tests treat Marconi as a black box. In other words, the API calls attempt to simulate an actual user. For example, unlike unit tests, the system tests do not use mock endpoints.

Running the System Tests

Setup a Marconi server. Refer to the Marconi README on how to run Marconi locally, or simply use an existing server.
System tests require the requests & robot packages. Run the following to install them: :
```
pip install -r tools/system-test-requires
```
cd to the marconi/tests/system directory
Copy etc/system-tests.conf-sample to one of the following locations:
```
~/.marconi/system-tests.conf
/etc/marconi/system-tests.conf
```
Update the config file to point to the Marconi server you want to run the tests against
If leaving keystone auth enabled, update system-tests.conf with a valid set of credentials.
Now, to run the sytem tests, simply use the pybot commands, e.g.:
Run all test suites: :
```
pybot queue/queue_tests.txt messages/messages_tests.txt claim/claim_tests.txt
```
Run test suites individually: :
```
pybot queue/queue_tests.txt
pybot messages/messages_tests.txt
pybot claim/claim_tests.txt
```
Note: pybot will generate report.html & log.html after the test run is complete.

Adding New Tests

See also the Robot user guide for more details on writing test cases.

Add a test case definition to an appropriate robot test case file: :

queue/queue_tests.txt
messages/messages_tests.txt
claim/claim_tests.txt).

Add test data to the test_data.csv in the same directory as the test case file you updated above (eg. queue/test_data.csv)
Add any validation logic you might need, to the following utility modules:
- corresponing *fnlib.py (e.g. queue/queuefnlib.py)
- common/functionlib.py (i.e., if the code can be used across multiple test suites)