turn down the metronome volume level or shut it off completely .....
or....
record it into a track as audio and use envelopes to lower the levels in the soft parts....
or....
use better headphones. I have used my noise reducing ear buds to record vocal takes and having a fairly loud can mix is a good thing to me. With headphones, that spill gets picked up by the Rode..... when I used my ear buds, the sound was just as good and just as loud and there was ZERO spill to the Rode condenser.
you gotta do what works best for you.