Separate the read/write SQL requests

Hi OpenStack community, I'm currently working on read/write database splitting in my OpenStack environment using this document: https://docs.openstack.org/large-scale/journey/configure/database.html#rever... Here is what I've done up to this point: - Configured HAProxy as recommended - Added the slave_connection parameter to all service configurations(nova.conf, cinder.conf, etc.) To test this, I enabled MariaDB's general_log on all nodes. However, upon checking the logs, most SELECT queries still reach the first database node instead of the other replicas. Did I miss a configuration step somewhere? Any guidance would be super helpful.