Resolve bow in bottom connecting skin - part 1

I approached Sling technical with the following:

I have noticed the undercarriage skin CF-SKN-003 is dishing upwards by about 3mm.

Note that the undercarriage has been fully shimmed with layered stainless steel sheet cut according to your pattern, packed in enough to achieve a pretty snug fit, and the undercarriage bolts fully torqued.  

This dishing was probably present from the moment it was rivetted, but it is possible it was made slightly worse when a small fraction of mm gap between the shims and the undercarriage was closed up further when the bolts were fully tightened.

My presumption is that the undercarriage skin CF-SKN-003 is supposed to transfer load from the front underside skin CF-SKN-004 to the centre fuselage bottom skin CF-SKN-002 all along its lateral extent. Now because this skin is dished it cannot do this job and these skin loads are being transferred to the undercarriage bolts.

My question is whether the dishing actually matters. If it does matter, one fix is to de-rivet along the aft edge of the undercarriage skin and allow it to relax flat. Some misalignment will then appear between the aft undercarriage skin holes and the centre fuselage bottom skin holes, etc, which should be a fraction of a mm. I then upsize this line of holes to 4.8mm absorbing the mismatch then re-rivet with 4.8mm rivets. All the relevant edge distances seem reasonable to allow this.

Sling's response was positive:

You can de-rivet and upsize if need be. Do you have an estimate about how big the misalignment would be? Would it be small enough to not have to go up a size in rivets?

I pointed out that:

I expect the misalignment to be about 0.8mm or so so upsizing to 4.8mm is feasible. The nature of the misalignments is such that the enlarged, recentered holes will be shifted slightly away from the sheet edges - the amount of metal between the hole edges and sheet edges will not be reduced.

I opened up the aft line of rivets and have since found that the misalignment is actually smaller than 0.8mm, more like 0.4mm. So the query raised above by Sling is even more relevant.

However, Sling did confirm that, given the above:

Then it is acceptable to upsize to 4.8mm.

I am still deciding what to do.