[tialaramex]

I have never done what you describe, and so I can't say "from my experience".

However, I would say that Job #1 is to identify what the end users (and the customers if the two are different) think this does. Reproducing the system they use/ paid for is the important goal of the rewrite, by definition if you can't fill that role you failed. Technical considerations (it would be nice to have Continuous Integration, it should be in C# not Java, it should run in the Cloud) are necessarily less important.

If you currently have little or no documentation this process produces valuable documentation and it concentrates you on what's important versus unimportant in a way that studying the existing software as a programmer does not. The trickiest code in the software could be so vital that without it your rewrite is useless or equally never used and you can't figure that out by staring at the program's code.

Once you have that high level you're in better shape to drill down and then I think you're in a better place to figure out if a rewrite is appropriate versus maybe you can refine the existing software.