The RFC for cover statements mentioned:

We can consider adding an option that would cause verification to fail if a cover property was unsatisfiable or unreachable, e.g. --fail-uncoverable.

We are now aware that some users want this option.

It'd be great if we could include this option as part of the "more granular expectations" discussed here. Let's say we had a language to describe expectations for regular properties and cover statements. Then, this expectation is essentially that all cover statements are satisfied.