AWS RDS: Managed Relational Databases, Multi-AZ & Replicas

AWS RDS

Amazon RDS (Relational Database Service) runs managed relational databases — PostgreSQL, MySQL, MariaDB, SQL Server, Oracle — so you keep familiar SQL while AWS handles the operational toil: provisioning, backups, patching, monitoring, and failover.

For most teams, RDS beats self-hosting a database on a VM. You trade some low-level control for automated durability and high availability, which is almost always the right call unless you have a specific reason to manage the engine yourself.

TL;DR

Managed relational databases — less ops, familiar SQL.
Plan Multi-AZ (high availability) and read replicas from the start.
Pool connections, especially from serverless clients.
Prefer RDS over self-hosting on EC2 unless you have a strong reason.

Quick Example

Provisioning a Multi-AZ Postgres instance is a single CLI call — AWS handles failover, backups, and patching:

Core Concepts

Managed engine — Postgres, MySQL, etc., with automated backups, snapshots, and patching windows.
Multi-AZ — a synchronous standby in another availability zone for automatic failover.
Read replicas — offload read traffic (engine-dependent).
Aurora — AWS's cloud-native, higher-performance MySQL/Postgres-compatible option in the same family.

Best Practices

Keep the database in private subnets; never expose it to the public internet.
Use least-privilege database users; enable encryption and enforce TLS.
Turn on Multi-AZ for production HA, and test restores (not just backups).
Pool connections (e.g. RDS Proxy / PgBouncer) for serverless and high-concurrency apps.
Monitor slow queries and add indexes; right-size storage and IOPS.

Scaling

See Caching and Database Replication.

Common Mistakes

Connection exhaustion from serverless

Skipping Multi-AZ

FAQ

RDS or self-hosting a database on EC2?

RDS for almost everyone — it automates backups, patching, monitoring, and failover, which are easy to get wrong by hand. Self-host on EC2 only when you need OS/engine-level control or a configuration RDS doesn't support, and you have the expertise to operate it.

What is Multi-AZ and do I need it?

Multi-AZ maintains a synchronous standby replica in another availability zone and fails over to it automatically if the primary fails. For production, yes — it's the main mechanism for high availability and zero-data-loss failover within a region.

RDS or Aurora?

RDS runs the standard engines (community Postgres/MySQL, etc.). Aurora is AWS's cloud-native, MySQL/Postgres-compatible engine with higher performance, faster failover, and storage that scales automatically — at a different price point. Choose Aurora for demanding workloads, RDS for standard ones.

How do I handle too many connections?

Use RDS Proxy or an application-side pooler. Relational databases (especially Postgres) allocate resources per connection, so large fleets of serverless functions can exhaust the limit. A pooler multiplexes many clients onto few backend connections.