The test {{test_validate_tally_on_openshift_cores_for_new_candlepin_facts }}uses randomize parameterization.

threads_per_core 2/3 of the times assumes a null value.

number_of_cpus have 0 as possible value, that probably represents an edge case. The fact that could not be picked 4/5 of the times is a bad practice.

We want to investigate if the test can be split depending on the expected behavior for the used values.

A white box boundary value analysis would help boiling down test cases.

A random number is fine but a range must be provided, so no "magic numbers" appear as parameters.

Acceptance criteria:

• Test(s) verifying that behavior run correctly in ephemeral.
• There is no pool of number from which parameters are picked.
• Edge cases are identified and covered by tests.