Adding support for supressing UNKNOWN results via --no-unknown #63
Conversation
| if len(filtered_messages) == 0: | ||
| messages_concat = 'OK' | ||
| if no_ok: |
There was a problem hiding this comment.
If the user passes --no_ok --no_unknown
UNKNOWN will be added to the output even if one was not removed.
| @@ -728,7 +730,7 @@ def test_print_results(check_docker, capsys, messages, perf_data, expected): | |||
| assert out.strip() == expected | |||
|
|
|||
|
|
|||
| @pytest.mark.parametrize("messages, perf_data, no_ok, no_performance, expected", ( | |||
| @pytest.mark.parametrize("messages, perf_data, no_ok, no_unknown, no_performance, expected", ( | |||
There was a problem hiding this comment.
You added a test parameter but you didn't update the data to match it. Have you used parameterized pytest tests before? It is pretty simple (and also a little ugly).
The string is a ordered list of names for all of the values in each of the tuples that follow. Each tuple in the list describes the setup of a single test case. For example
def test_func(one, two)
...
This executes does something like this
for params in ((1,2),(11,22)):
one, two = params
test_func(one=one, two=two)
Can you add the missing entries to the tuples and also add tuples that exercise the code you are changing.
Don't worry if you break something, I can help you out. Also my dev documentation should help you get the test running on your machine.
I have a use case where a bunch of containers that I query via --containers all do not have healthchecks defined, this results in an UNKNOWN return. This PR adds support for supressing these messages.
For the test class, I am not entirely sure if I am doing this right, so any feedback/corrections welcome.