Since we recently we recently had 2+ hours of downtime, I will allow it. /s
TBH, I would do the downtime because if there is a problem, you will be able to fallback quickly.
You want to failfast.
If you try the migration path, there could be an undetected ptoblem ocuuring that is not discovered quickly, and the remediation could end up being a bigger mess than 2 hours of downtime.