As Splynx evolves to be the core BSS/OSS for WISPs it has become necessary for Splynx to natively support active-active AAA and DB backend.
Especially when many carriers now provide MPLS VPLS and other layer 2 services across physical /geographical locations which makes sense for medium to large WISPs to have DR or second NOC centers to hosts replicated business support systems.
I am aware splynx has a non-native active-passive setup documentation using Proxmox but we need to start planning for horizontal scaling where two or more AAA servers operate concurrently with the same information served from a physically/geographically decentralized location to ensure 99.99999% uptime.
Some years back we used https://severalnines.com/blog/deploying-active-active-freeradius-cluster-mysql-ndb-or-galera as a reference to build a custom AAA and DB service. Perhaps this could be modified by the splynx development team to provide native active-active splynx applications services not only for recovery but also for load balancing.