I count the "glue it with existing infrastructure" to be higher cost than doing it from scratch. It was one feature that I definitely knew regarding Nomad, as one or two people who used it did chime in years ago in discussion, but for various reasons that might not be applicable to everyone I consider those unnecessary complication :)
Depending on how big the infrastructure is and how long you want to migrate over... usually there's not enough resources to "redo all from scratch", millions of LoC are already in production, people who owned key services are no longer in the company, other priorities for business exists other than have what you have already working in k8s..
The context was rather different (home setup), but all that you mention can be used as arguments Noth for and against redo, basing on situation in company, future needs, etc.
I have actually done a "lift and shift" where we moved code that had no support or directly antagonistic one to k8s because various problems reached situation where CEO said "replace the old vendor completely" - we ended up using k8s to wrestle with the amount of code to redeploy.
