You mentioned separate data centers and geographic redundancy. I'm not exactly sure how you're defining those two, because geographic redundancy could be a separate data center in the same region, or completely different regions. I'm assuming you're talking about separate data centers in the same region though.
I don't have a lot to input about that, except that you are going to achieve maximum performance if all components ArcGIS Enterprise is using (not just Esri, but DB servers, etc.) are all located in the same data center. At the end of the day, I think it would be up to the end user/organization to determine how much latency is "acceptable."
I can also speak from personal experience that having components spread across multiple regions (like AWS East and West) will be terrible for performance. You might be able to get away with it if they're in separate data centers in the same region, but it's certainly going to be bad if they're in different regions.
