De-duplicate redundant feature tests
We write a lot of tests (approx. 1000 per month), but we almost never remove redundant tests.
We should clean up such tests.
One first step could be to name feature tests according to https://docs.gitlab.com/ee/development/testing_guide/best_practices.html#system--feature-tests and then it will become more obvious what we're testing multiple times.
We should also investigate if we can discover redundant tests programmatically. We can use the "Avg. hits / line" metric here: https://gitlab-org.gitlab.io/gitlab-foss/coverage-ruby/#_AllFiles but I'm not sure this is very useful.
Edited by David Dieulivol