[QA] Add a new Test::Sanity::Failing scenario that always fails

What does this MR do?

The Test::Sanity::Failing scenario should always fail. Its purpose is to check that our QA test setup works as expected. We need to do that because a RSpec run could skip all the tests and exit with a 0 code, and hide the problem.

By having a test that is always supposed to fail, we can run it and expect that its exit code isn't 0 in a CI job (in gitlab-qa, quality/nightly, and quality/staging).

Right now, I expect the sanity test to fail in gitlab-qa since we have a problem with our orchestrated scenarios not being run (see #50963 (closed), fixed by https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/21468).

What are the relevant issue numbers?

Closes gitlab-qa#28 (closed). Relates to gitlab-qa!171 (merged).

Does this MR meet the acceptance criteria?

• [-] Changelog entry added, if necessary
• [-] Documentation created/updated
• [-] Tests added for this feature/bug
• Conforms to the code review guidelines
• Conforms to the merge request performance guidelines
• Conforms to the style guides
• Conforms to the database guides