Lesson 3 of 16

AWS RDS for PostgreSQL: The Managed Baseline

Applies to PostgreSQL 13–17 Last reviewed Jun 2026 Grounded in source

The one thing to understand first

Amazon RDS for PostgreSQL is the purest example of the Postgres-on-a-disk family: it is genuine community PostgreSQL running on an EC2-class instance with an Amazon EBS volume underneath, wrapped in AWS automation. If you understand how you would run PostgreSQL on a Linux box with a separate data disk, you already understand RDS — AWS has just automated the operations around it.

Storage and the single-writer model

Your heap and WAL live on one or more EBS volumes attached to the instance. RDS supports storage autoscaling (the volume grows when it nears full) and lets you choose the volume type and provisioned IOPS. Because it is ordinary PostgreSQL, there is exactly one primary that accepts writes; everything else is a copy fed by replication.

High availability: Multi-AZ in two flavours

Multi-AZ instance. RDS keeps one standby in a second availability zone, kept current by synchronous physical replication. The standby is not readable — it exists purely for failover. If the primary fails, RDS flips a DNS <a class="sev1-termlink" href="https://thesev1database.com/glossary/tuple/">record to the standby, usually within a minute or two. You get durability and automatic failover, but no extra read capacity from the standby.

This is a Pro lesson

Get every Learning Pathway and cookbook recipe — grounded in PostgreSQL source code, with diagnostics, fixes, and prevention for each topic.

Continue this lesson to learn:

  • Read replicas, backups, and recovery
  • How it differs from self-managed PostgreSQL
  • What people miss
  • Key takeaway
  • All 36 Learning Pathway lessons
  • 170+ cookbook recipes
  • Source-grounded diagnostics & fixes

Secure checkout Cancel anytime Source-grounded

Keep going

Related & next steps

Was this helpful?

← Back to 06 — Managed PostgreSQL in the Cloud: Architecture & Trade-offs