I'm setting up Jira, Confluence & Bitbucket environments with external PostgreSQL database. Database will have few replicas for redundancy and high availability purposes. Replicas can also serve SQL queries, but only read queries. So I'm trying to figure out if there is a way to separate read and write queries, so I could point read queries to the replica node and write queries to the master node. This way I could load balance SQL and make better use of the otherwise idle replica node.
