APU: No Pro = No ECC. Make that a Ryzen Pro 5650G or 5655G.
The CPU is the less important part of NAS, really. But historically, ECC (and IPMI) have been strong reasons to go with Intel: No official ECC support with Ryzen, and not many server-grade boards with IPMI.
Back to the first question, having two NAS, the primary replicating to the secondary is the first step into a backup strategy. Then come “off-site” and “off-line”.
What is the rationale to have two NAS with two different sets of data and different compliance with “best practices”, rather than consolidating everything into one box? What is the point of having a GPU in the “most important” ECC-enabled NAS… if videos are played from another NAS, elected because it has an HDMI output?
I could understand, for instance, having a primary NAS for the most important data with an all-SSD pool to be silent and stashing big bulky HDD storage elsewhere, away from ears. But here it seems that capacity requirements would be on the UGREEN NAS that is in the living room next to the TV and that noise is not a concern for the primary (1U/2U rack).