The 3-2-1 backup strategy is a best practice that suggests having three total copies of your data, two of which are local but on different devices or media, and one copy offsite. In the described scenario, although there is an offsite component (SOBR with AWS S3 Performance Tier), and a local hardened immutable repository, there is no indication of a second local copy on a different device or media. This means that there is only one local copy of the backup data and one offsite copy. Therefore, the setup does not meet the 3-2-1 best practices requirement, which would involve having at least one more local copy on different media or device.

Explanation
According to the Veeam Backup & Replication user guide1, you can use a long-term retention policy (GFS) to keep backup files for longer periods oftime. GFS stands for grandfather-father-son and refers to a backup rotation scheme that allows you to create weekly, monthly, quarterly and yearly restore points12.
Therefore, if you want to be able to restore the VMs to any given day within two months and any given week within six months, you should use a GFS policy with 62 days of retention and 26 weekly GFS restore points.
The simplest configuration would be an immediate copy (mirroring) mode with these settings (option B).

Given the proxy load of 68% and all other loads (source, network, target) at 0%, the bottleneck is at the proxy. To decrease the backup processing time and accommodate the growth of the environment, adding additional Windows or Linux machines as backup proxies would help distribute the load. This would enable parallel processing and reduce the time taken to complete the backup jobs, thereby helping to meet the RPOs.