Without seeing an uploaded sample or something, it’s hard to speculate, but there’s definitely nothing wrong with rebuilding the surface to something that is still within tolerance.
‘Rebuilding’ is a common part of most of my workflows. I usually prefer homogeneous isocurve density, but it all depends on what’s necessary for the job at the time.
Sometimes fitting a surface really tight is inevitable if you want the single surface to conform to other adjacent geometry.
For example: (sweep 2 rail) Modeling Problem/question - #16 by lander
But depending on the ‘rules’ those situations are also capable of being mitigated with further design revisions – if necessary.