poor execution from cell 2 to cell 1 has nothing to do with congestion, but probably more with a broken TRX in cell 1
what is the call setup success rate of cell 1 ?
more precisely, what is the TCNAFLR (tch normal assignment failure rate) ?
if cell 1 is working perfectly, then you should check the radio conditions in the area where MS are most probably doing the HO from cell 2 to cell 1. Check interference, or obstacles, that could stop the radio link suddenly. However, with 7% success rate, I would assume it’s more of a TRX hardware problem than a radio problem.