Jim Adney writes:

Interesting. Across one cell, it drops from 5.4V to 2.8V, when initially
plugged in, then after the UPS chirps and dies, it gradually rises again. Not
sure what's going on here.

That makes me think that the cells are in there backwards.

The cells are in the packs backwards, or the packs are in the UPS backwards? I
don't see how the packs could be backwards, they are hooked up exactly as the
old ones came out. The +/- leads are too short to reach to the wrong
respective terminals anyway.