The current setup of the hub is very centralised. A more distributed hub would help resolve issues like latency (which can be a problem during deployment, see example issue). Next to this, it's a single point of failure. Even tough we now have a secondary instance which could be used to read in case the primary goes down, it's still up to users & software to know & use the secondary and writes are also not supported while the primary is down.
The current setup of the hub is very centralised. A more distributed hub would help resolve issues like latency (which can be a problem during deployment, see example issue). Next to this, it's a single point of failure. Even tough we now have a secondary instance which could be used to read in case the primary goes down, it's still up to users & software to know & use the secondary and writes are also not supported while the primary is down.