You find out how good your RDS backups are the day someone asks for one customer's rows from last Tuesday, and the database lives in account 212 of 400. (Nobody ever asks during a slow week.)
With one account, native snapshots and point-in-time recovery cover you. Once you're running hundreds of accounts across several Regions, and maybe a second cloud, the hard part moves from taking backups to proving every database has one and pulling back the exact data you need.
I've evaluated these tools at both ends of that range, so I sorted these RDS backup solutions by how far each one scales before the gaps show.
Here's where I'd start:
The best RDS backup solution for enterprise AWS teams is Eon, a cloud data protection and recovery platform that discovers new RDS and Aurora databases across hundreds of accounts, applies policy without tagging, and restores down to the rows a query matches.
Keep Amazon RDS automated backups on as the base layer, with manual snapshots and cross-Region automated backups for extra copies. AWS Backup adds org-wide policy, cross-account copies, and Vault Lock, while N2WS orchestrates snapshots from your own account.
Veeam fits VM-heavy estates, Druva brings a SaaS-delivered vault, Commvault (Clumio) adds long-retention tiers, and Rubrik suits security-led teams with a broader estate.
Where native RDS backup stops scaling
Native RDS backup is configured per instance, which works until the estate grows faster than anyone's spreadsheet. Retention defaults to 1 day when an instance comes from the API or CLI and 7 days from the console, so two databases built by two pipelines can carry very different safety nets.
Then there's the ceiling. Automated backups and point-in-time recovery top out at 35 days, so audit-length retention means manual snapshots, which default to a quota of 100 per Region, or AWS Backup snapshots, which don't count toward it.
The third gap is location. Snapshots stay in the same account as production unless you copy them out, which puts the database and its recovery points behind the same set of credentials.
Native backup still belongs in every plan. At enterprise scale, the question grows from "is it on?" to "is it on everywhere, kept long enough, and stored where a compromised account can't reach it?"
1. Eon: a fit for hundreds of accounts across more than one cloud
I'd put this platform first for a growing AWS Organization, because it answers the question native tools leave to you: what's still unprotected?
Its Cloud Backup Posture Management (CBPM) continuously discovers resources as environments change, classifies them (PII, PHI, and financial data), and applies backup policy without manual tagging.
It connects through a single read-only role with no agents or appliances, which matters when you're onboarding account 400 (and 401 next week).
PDI Technologies, a convenience-retail and fuel software company, runs about 600 AWS accounts and saw success on its 15-minute-RPO backup jobs climb from about 80% on a hand-built Lambda workaround to around 99.95%.
For RDS, it complements the AWS-native layer you already run. It can manage AWS-native point-in-time recovery for RDS and Aurora across accounts and Regions, restoring to any second while the recovery points stay in your own account.
It then extends that layer with long-retention snapshots in an immutable, logically air-gapped vault in a dedicated account, with time-bound access, so a stolen production key doesn't reach recovery data.
The company recommends a dual policy, such as 7-day PITR plus a snapshot policy for long retention, compliance, and database exploration. Native PITR caps at 35 days, while snapshot retention runs as long as policy needs.
Restores go as narrow as you need: the instance, one database, one table, or only the rows a SQL query matches. That lets you roll back the exact records hit in minutes while later good writes stay put.
Backups are verified by fingerprinting every table, then periodically restoring to a temporary instance and comparing the fingerprints.
Ransomware detection (add-on) correlates multiple independent signals (such as row-count drops, schema changes, and ransom notes) into one weighted score, and detection for databases is in early access.
Coverage reaches Cloud SQL and Azure SQL, and RDS and Aurora can be protected into a Google Cloud vault you can query and restore records from during a cloud-level incident, while a full restore returns the database to AWS.
Pricing is usage-based, billed per GB per month on protected storage, with no tier ladder and flexible spending commitments. The company cites up to 50% lower cost through global deduplication, and there's no public list price.
The platform's on-prem support is limited, so an estate that's still mostly on-prem needs a separate tool for the data center. Its customer reference list is also shorter than the incumbents can show.
2. N2WS: a fit for AWS estates that want DR orchestration inside their own account
N2WS, which now brands itself N2W, is a Linux-based virtual appliance you launch from an AMI in your own AWS account. It uses AWS APIs to manage snapshots of RDS databases (regular and custom) and Aurora clusters, including Aurora Serverless.
Its strength is cross-Region and cross-account DR for RDS, with flexible schedules that the pricing page says can run at a 60-second interval.
Two documented catches apply at scale. Snapshots encrypted with the default KMS key can't be shared across accounts, and N2W's docs limit its cross-account RDS copies to the DR account in the original Region.
For longer retention, it can copy RDS to an S3 repository protected with S3 Object Lock, using AWS's snapshot export to Parquet. That path covers MySQL and PostgreSQL only, and Aurora is excluded from repository copy.
Pricing is the draw for big fleets. As of October 2026, N2W charges a flat monthly subscription based on protected instances, so the bill doesn't move with data volume.
Standard costs $249 a month for 20 instances and 3 accounts, while Advanced (50 instances, 10 accounts) and Enterprise (200 instances, 20 accounts) are quote-based. Enterprise+ is for 1,000+ instances across 100+ accounts, and every plan offers a 30-day free trial.
Aurora gets one licensing wrinkle worth modeling: cluster storage counts toward the license in 10 GiB increments.
The tradeoff is ownership, because the N2W server is yours to run, patch, and upgrade. Upgrades ship as a new AMI that wipes OS-level changes, and RDS backups stay native snapshots unless you add the S3 copy.
3. AWS Backup: a fit for any estate already organized under AWS Organizations
AWS Backup is the policy layer most enterprise teams keep, and for good reason. From a single management account in AWS Organizations, you can manage backup policy across member accounts.
For RDS, it creates manual DB snapshots through the same create-db-snapshot operation RDS uses, so those snapshots share the instance's incremental chain.
Continuous backup gives point-in-time recovery with 1-second precision for up to 35 days, and RDS can be restored up until the most recent 5 minutes of activity. Snapshot rules can run as often as hourly and keep copies for up to 100 years.
For isolation, cross-account copies move snapshots into another account in the same Organization. Vault Lock in compliance mode makes a vault immutable to every user and to AWS once a grace period of at least 72 hours ends, at no extra charge.
Two scale caveats matter. Copies to another account or Region are snapshots, so the PITR window stays in the source account. Within the RDS family, logically air-gapped vaults accept encrypted Aurora clusters only, so RDS DB instances can't go into one.
Pricing is pay-as-you-go with no minimum fee. As of October 2026, RDS snapshots in AWS Backup bill at $0.095 per GB-month in us-east-1, Aurora cluster snapshots at $0.021, and restores of both are free.
It also runs scheduled restore tests for RDS and Aurora and audits policy compliance through Backup Audit Manager, which is exactly the evidence an auditor wants to see.
The limit shows up at restore time: every restore builds a new resource, and there's no item-level restore for RDS or Aurora. Recovering a few rows still means standing up a copy and moving data back by hand.
4. Veeam: a fit for VM-heavy estates that also run RDS
Veeam's AWS protection now lives inside Veeam Backup & Replication 13 as the Veeam Plug-in for AWS, which adds backup appliances to the console many teams already run for their VM estates.
For RDS, it works two ways. It can create cloud-native snapshots of RDS instances and Aurora clusters and replicate them to any Region in any account, or take image-level backups of SQL Server and PostgreSQL (versions 12 through 17) into Amazon S3.
Those image-level backups are the interesting part, because they allow database-level restores for SQL Server and PostgreSQL. A worker instance spins up in the production account for each job and is removed when the session ends, with no agent software inside the database.
As of October 2026, the plug-in is licensed per protected instance, and each RDS resource counts as one. A free edition protects up to 10 instances on appliances not managed by a Backup & Replication server, minus PostgreSQL image-level backups.
Paid use runs on a Veeam Universal License subscription. Veeam Data Cloud for AWS is the SaaS route, with all-inclusive pricing and no number on the US page.
MySQL, MariaDB, Oracle, and Aurora get snapshot protection only, so database-level restore is limited to the two engines above.
Scale is where the plumbing shows. Workers need VPC DNS resolution, a public subnet with an internet gateway or private endpoints, and two temporary security groups per job, and AWS, Azure, and Google Cloud are separate plug-ins, each with its own appliances.
5. Druva: a fit for SaaS-first teams that want no infrastructure to run
Druva comes from endpoint and SaaS protection and has extended into AWS with two layers. CloudRanger handles snapshot orchestration (scheduling, retention, cross-Region and cross-account copies, and DR plans).
RDS Airgap is the SaaS layer, storing backups in Druva's own cloud, decoupled from your AWS Organization. That separation appeals to teams that don't want to run any vault infrastructure themselves.
On Enterprise and Elite plans, Druva also markets Data Lock to prevent alterations and deletions, and it says cross-Region copies carry no egress fee.
Cluster support needs a careful read. Druva's own AWS support matrix (dated October 2025) says RDS Cluster backup and restore is not currently supported, while its RDS Airgap guide (updated May 2026) says you can protect RDS databases and RDS clusters.
Storage rules map accounts and Regions to Druva Cloud storage, and once a storage Region is set, existing backups can't move to a new one. RDS Airgap backups also can't follow an Azure storage rule, so they stay in AWS.
As of October 2026, Druva's public cloud plans (Business, Enterprise, and Elite) are priced per TB per month after deduplication, with a free trial and no published dollar figures.
Restore granularity is the other thing to check, since the support matrix marks file-level recovery for RDS as "NA" and Druva documents no table or record restore.
At scale, two limits bite. Airgap copy isn't supported for RDS instances encrypted with AWS managed keys (Druva recommends customer managed keys), and backups fail when an account passes its manual-snapshot quota (100 per Region by default).
6. Commvault (Clumio): a fit for regulated estates with long retention rules
Commvault covers RDS two ways. Commvault Cloud backs up and restores RDS instances across multiple accounts and Regions, and Clumio, a Commvault company, is its AWS-native line.
Clumio puts in-account automated snapshots and logically air-gapped backups of RDS instances and Aurora clusters under a single policy. SecureVault Standard targets retention under three months, and the Archive tier holds logical copies kept longer.
Archive adds documented record-level retrieval for RDS: you can pull individual records from the backup and load them into a database after a thaw period of up to 48 hours.
As of October 2026, Clumio lists RDS SecureVault Standard at $0.14 per GiB-month with free restores (data transfer out fees still apply). Archive is $0.01 per GiB-month, plus a $0.01 per GiB-month early deletion fee on data kept under 12 months and $0.025 per GiB restored, and Aurora Standard is $0.030 per GiB-month.
Engine coverage runs from Aurora MySQL and Aurora PostgreSQL to RDS for Oracle SE2 and Enterprise Edition. Db2 isn't listed, and Archive or record retrieval doesn't cover Oracle Enterprise editions.
Logically air-gapped backups have a minimum RPO of 4 hours, and the Archive tier has no full-database restore and takes up to 48 hours to start. Clumio also builds on RDS manual snapshots, so AWS snapshot quotas still apply.
7. Rubrik: a fit for mixed on-prem and multi-cloud estates under one policy engine
Rubrik Security Cloud is the pick when RDS is one workload among many and the security team wants one policy engine across on-prem and multiple clouds. Rubrik calls itself the Security and AI Operations Company.
For RDS, a central console applies SLA policies to instances across accounts and Regions, and setup creates a cross-account IAM role through CloudFormation without long-lived keys.
Rubrik's RDS data sheet, dated September 2023, lists MySQL, SQL Server, Oracle, MariaDB, and PostgreSQL, and Rubrik protects transaction logs for point-in-time recovery. Its RDS page says it restores RDS instances to an exact point in time within minutes.
There are two deployment models. Customer-hosted runs Rubrik's compute in your own AWS account, so your team owns the networking and IAM (and Rubrik itself warns that environments drift).
Rubrik Managed Cloud Protection runs compute and the vault in Rubrik's environment, with Object Lock on every backup by default. It has narrower configuration options today and moves data outside your AWS organization by design.
Rubrik's security-led cyber recovery is strongest where it can scan files, because file-based scanning reads storage files and a managed database like Aurora or RDS never exposes its contents as files to scan.
As of October 2026, Rubrik has no public price list, so expect a sales quote.
Rubrik's RDS data sheet dates from September 2023 and doesn't name Aurora, although Rubrik announced Aurora protection separately in 2023. Rubrik's original RDS protection rides on native snapshots, so the AWS default quota of 100 manual snapshots per account per Region applies to it.
8. Amazon RDS automated backups: a fit for every estate, as the base layer
Amazon RDS automated backups belong underneath every tool above, because they're free up to your provisioned storage and need almost no setup. RDS snapshots the whole instance, keeps later snapshots incremental in Amazon S3, and uploads transaction logs every five minutes.
Instance retention runs from 0 to 35 days, and 0 turns automated backups off. Aurora backups are continuous, can't be disabled, and the latest restorable time is typically within 5 minutes of the current time.
Manual snapshots fill the retention gap because they don't expire, though each Region allows 100 by default. Cross-Region automated backups copy snapshots and transaction logs to a second Region, with a default quota of 20 per account, and Multi-AZ DB clusters aren't supported.
Since September 2025, you can also copy RDS and Aurora snapshots across Regions and accounts in a single step. Aurora MySQL adds Backtrack, which rewinds a cluster in place within a 72-hour window if it was enabled when the cluster was created or restored, although it can't target a single table.
As of October 2026, backup storage up to 100% of your total database storage in a Region costs nothing extra, and RDS charges $0.095 per GiB-month beyond that in us-east-1. Aurora backup storage outside the free allowance runs $0.021 per GB-month.
The limits are structural: retention tops out at 35 days and every restore builds a new instance. Backups also sit in the production account until you copy them out, so keep this layer on everywhere and build the rest on top of it.
A rollout order that works
This sequence keeps each step small enough to finish before the next audit cycle starts.
- Check retention on every instance first, because databases created through the API default to 1 day of automated backups.
- Set AWS Backup policy from the management account so every member account inherits a baseline.
- Move long-retention copies out of production accounts and into a Vault Lock vault or a separate account before you need them.
- Add record-level restore and restore verification for the databases where one bad write costs the most.
- Schedule restore tests for RDS and Aurora, and keep the reports where your auditors can find them.
- Extend coverage to other clouds once AWS coverage is boringly consistent (boring is the goal here).
Make every new account protected before anyone asks
The enterprise teams I'd trust with RDS share one habit: new accounts land under policy on day one, and somebody has tested a restore recently.
Before your next quarterly review, pick five databases at random and check three things: a policy covers each one, a copy sits outside its production account, and someone restored from it in the last 90 days. Any "no" is your first project.
FAQ
How do you back up RDS across hundreds of AWS accounts?
Start with AWS Backup policies managed from a single management account in AWS Organizations, then add a tool that discovers new databases and applies policy automatically so coverage doesn't drift as accounts are created.
Does N2WS run inside my AWS account?
Yes, the N2W server is a Linux-based virtual appliance you launch from an AMI in your own AWS account, and it uses AWS APIs to manage RDS snapshots. You run, patch, and upgrade it yourself.
Do cross-Region copies of RDS backups keep point-in-time recovery?
In AWS Backup, cross-Region and cross-account copies are snapshots, so the PITR window stays in the source account. RDS's own cross-Region automated backups copy both snapshots and transaction logs, although Multi-AZ DB clusters aren't supported.
Do RDS backup tools cover Aurora clusters?
Most do with caveats: AWS Backup, N2WS, Veeam, Commvault (Clumio), and Rubrik all document Aurora support. Druva's AWS support matrix (dated October 2025) lists non-serverless Aurora but also says RDS Cluster backup and restore is not currently supported, while its May 2026 Airgap guide says they're protected.
