If I understand correctly the net effect of the change means log_slave_updates option is now always enabled, on a MariaDB 10.3.28 3-node Galera cluster with binary logging enabled, I now see all writes to all nodes appearing in the binary logs on all nodes which is a change in behaviour. Unfortunately as the change came from a merge from the 10.2 branch it has taken me a while to track down as I have seen an impact on disk IO and disk space on my cluster nodes since upgrading from 10.3.9 to 10.3.28. I am concerned about the extra demands on Galera cluster nodes, and is there any way to achieve the previous behaviour?

I think I may have jumped the gun, the change here is about always enabling the binary log irrespective of the log_slave_updates setting, as I enable the binary log explicitly in my configuration it should have no effect. However, there is a change of behaviour where all nodes are logging all updates to all binary logs in 10.3.28, I'll see if I can track down where that change happened and raise a new ticket.