The library / Relational databases

RDS availability is not a backup strategy

Separate failover, read scaling, and point-in-time recovery before choosing your managed database configuration.

Start with the failure

Different features solve different problems. A failed compute instance, an accidentally deleted table, and a reporting workload that overwhelms the primary are not the same incident.

List the failure you need to survive before selecting a deployment option.

Availability keeps service running

A deployment with a managed failover capability can improve availability when an infrastructure component fails. The supported topology, reader behavior, and failover details vary by RDS engine and deployment type.

Applications still need sensible connection handling and retries. Existing connections may not survive a failover.

Recovery gets earlier data back

Replication can faithfully copy an unwanted change. If an application deletes the wrong records, a standby may receive that deletion too.

Backups and point-in-time recovery address a different requirement: reconstructing a database at an earlier state. Recovery procedures must include validation and application reconnection, not just creating a restored instance.

Read scaling has another purpose

Read replicas or readable instances can serve some read workloads, subject to the chosen engine's consistency and topology behavior.

Do not assume every standby is readable. In particular, distinguish RDS Multi-AZ DB instance deployments from Multi-AZ DB cluster deployments and Aurora architectures.

Make a recovery worksheet

Record the acceptable data-loss window, acceptable downtime, backup retention, encryption dependencies, restore permissions, and the person responsible for declaring a recovery complete.

Test the full restore procedure in a controlled environment. Check both data correctness and the time needed to make the application usable again.

Review costs and lifecycle

Additional instances, storage, backups, and cross-region transfers can create charges. Check the current service documentation and estimate costs for the exact configuration and region.

This guide does not create infrastructure. A hands-on deployment should include explicit teardown instructions and a decision about which backups to retain.

Keep exploring

Go deeper with the original documentation.

Official documentation
D
DBMinutes Editorial

Practical explanations of database systems, cloud services, and the engineering decisions between them.

AI-assisted content · Our editorial process

Keep the curiosity going.

Back to the library
A little learning goes a long way

Make room for a few good minutes.

Join the list for practical guides, thoughtful comparisons,
and ideas worth bringing to your next project.

Find your next answer

Search concepts, tools, and practical guides.