Make sure the measurements accurately reflect what you care about. What is code quality for you? Do you care about customers not encountering bugs? (All bugs? Just critical?) Do you care about development speed not being reduced gradually? Do you care about estimates matching actual time spent, so you can plan ahead? If you don't actually care about how many lines of code are committed to the git repository, then obviously don't measure it!
And if you can't measure something for each team member individually, make the whole team equally accountable for the end result. They will figure it out themselves and tell you who's bringing the team down. Not just that, once they know what you really want, they'll also be able to suggest how to measure it better, or tell you what you can do to let them achieve greater results. (E.g., faster PC, bigger monitor, fewer distractions, no scope changes, better test environment, specific training, etc.)