0.05mm is borderline. On our 200T Komatsu we hold 0.02mm parallelism spec and I would not let 0.05 slide (no pun intended). raj_pm is right about checking at multiple stroke positions — that is the key diagnostic step most people skip.
Here is what we do. Four dial indicators on magnetic bases, one at each corner of the bolster pointing up at the slide face. Jog the slide down to BDC in 5mm increments and record all four readings at each position. Plot it on a spreadsheet. If the error is consistent through the stroke, your frame alignment is off — could be foundation settling, could be a cracked tie rod on an older C-frame. If it only shows up near BDC, it is gib wear and you can fix it with adjustment or regrinding.
For the actual correction: loosen the gib lock screws, adjust the set screws on the high side in 1/4-turn increments, re-torque the locks, and re-measure. It is tedious but do not rush it. We usually spend 2-3 hours on a full parallelism correction. And always re-check after running 500 parts at operating temperature — thermal growth can shift things 0.01-0.02mm.