Multiple reservations on a single compute resource

I have multiple tenants (A and B)

with multiple BGs inside each tenant (A_BG1, A_BG2, B_BG1,B_BG2)

I have a compute resource FG_ABC

If I need to provision workloads from both BGs (A_BG1 and B_BG1), I'll need 2 reservations; If I need all 4 BGs to provision workloads I'll need 4 reservations.

**Check my math here** Just want to make sure.

Reservations work off exact numbers (GB/memory, GB/storage,etc).

1.  How do you calculate these resource reservations?  Do you just say A_BG1 gets 75% (which calculates to ~75GB) and B_BG1 gets 25% (which calculates ~25GB).

2.  If you add 100GB to the FG (say we add a host to a cluster), do you have to manually reconfigure those reservations to include this extra 100GB?

Just looking for feedback on how I should be reserving resources.  Thanks.