When running maestro test, we leverage before_maestro_tests and after_maestro_tests hooks to setup and teardown our test environment but the after_maestro_tests is not run if one of the maestro test fail which leave the environment in an instable state