PostgreSQL 18 on Amazon Aurora and Amazon RDS: Enhanced Security, Monitoring, and Developer Features

PostgreSQL 18 improves security, operational monitoring, and developer productivity on Amazon Aurora PostgreSQL and Amazon RDS for PostgreSQL. Key changes include the deprecation of MD5 password authentication support and the addition of new statistics metrics that track parallel worker activity.

MD5 authentication will continue to work for the time being in PostgreSQL 18 but is scheduled for removal in future releases. The default password_encryption value in AWS environments is already set to scram-sha-256, but roles using existing MD5-hashed passwords must reset their passwords to migrate to SCRAM-SHA-256.

The new md5_password_warnings GUC parameter is enabled by default and causes CREATE ROLE and ALTER ROLE to log deprecation warnings in the engine log when passwords are stored using the MD5 method. These warnings can be used to identify roles that are still using MD5 passwords.

Two columns showing parallel worker activity have been added to pg_stat_database and pg_stat_statements.

• parallel_workers_to_launch: The number of parallel workers requested by the planner for execution
• parallel_workers_launched: The number of parallel workers actually launched

The difference between these two values allows you to check if parallel workers were insufficient due to limits such as max_parallel_workers or max_worker_processes. pg_stat_database provides cumulative status across the entire database, while pg_stat_statements identifies individual queries affected by worker shortages.

For example, if the planner requested 66 workers but only 20 were launched, 46 workers were not executed. If this situation persists, you can increase the parallel worker limit or disable parallel processing for parts of the query to reduce resource contention.