Is there a way to set a low minimum sample size for an experiment, just to check that an experiment is receiving the expected data?

yes, you can do this per metric