Postgresql Non-transactional and Transactional connection

[JOSEGEOVO]

How can I configure the DB connection to use Aurora proxy? I want to specify when to use a reader instance and when to use a writer instance. Not sure if we can specify two connection strings (One for reader and another for writer)

[djvirus9]

For the v4 transactional/non-transactional pool configuration, there is one database URL, not separate writer and reader URLs. Both pools are constructed from the same alpine.database.url. The pool-specific tx/nontx properties adjust sizing and timeouts; they do not select different database endpoints. This is visible in the Alpine 3.1.2 connection factory used by Dependency-Track 4.12.7.

To connect that deployment through RDS Proxy, point the existing URL at the proxy's read/write endpoint, keeping your configured database name, credentials, and TLS settings. For example, the host portion becomes the proxy endpoint rather than the Aurora instance endpoint.

That gives you proxy connection management, but it does not offload reads automatically. AWS documents that the default/read-write proxy endpoint sends requests to the writer. A separate read-only proxy endpoint sends requests to readers, and the application must explicitly use that endpoint for the appropriate workload.

Consequently, creating a reader endpoint in AWS alone will not make this Dependency-Track configuration use it. Its two pool names describe transaction handling, not a supported read-replica routing interface. Pointing its only URL at a reader endpoint would also send required writes there, where they cannot succeed. The v4 database documentation describes the pool controls available without changing application code.