AWS RDS vs Azure SQL Database: Which to Choose (2026)

Quick verdict

Short answer

Choose Amazon RDS if your applications run on AWS, if you need an engine other than SQL Server (PostgreSQL, MySQL, MariaDB, Oracle or Db2), or if you are moving an existing SQL Server instance that relies on SQL Server Agent, several databases per instance or native .bak backups. Choose Azure SQL Database if you are on Azure and building a new SQL Server application that fits within a single database, and you want serverless compute, Hyperscale and a free offer without managing an instance. For SQL Server lift-and-shift into Azure, compare RDS for SQL Server with Azure SQL Managed Instance rather than with Azure SQL Database.

How we know: This comparison is research-based: supported engines, editions, high availability, backup, maintenance and pricing models were checked against the AWS RDS product and pricing pages, the Amazon RDS User Guide and Microsoft Learn in October 2026. We have not run performance or cost tests, and nothing here is a benchmark.

Amazon RDS (Amazon Relational Database Service) is AWS's managed relational database service. The Amazon RDS User Guide lists six engines: IBM Db2, MariaDB, Microsoft SQL Server, MySQL, Oracle Database and PostgreSQL. The AWS RDS product page also counts the Amazon Aurora family (Aurora MySQL-Compatible, Aurora PostgreSQL-Compatible and Aurora DSQL) under the RDS banner, but Aurora has its own user guide and architecture; see Amazon RDS vs Amazon Aurora. You create a DB instance of a chosen class (db.m*, db.r*, db.t* and so on), and AWS handles backups, patching, failure detection and recovery. You get no shell access to the host. RDS Custom is a separate variant, for Oracle and SQL Server only, that does give you access to the operating system and database environment for applications that need it.

Azure SQL Database is a fully managed database service in Azure that runs the latest stable SQL Server Database Engine. It is database scoped: you create single databases or elastic pools on a logical server, with no instance and no operating system to manage. It offers a vCore purchasing model (General Purpose, Business Critical and Hyperscale tiers, with serverless compute in General Purpose and Hyperscale) and an older DTU model. It is one member of the Azure SQL family, alongside Azure SQL Managed Instance (a managed instance with near full SQL Server compatibility) and SQL Server on Azure VMs. For other engines, Microsoft sells separate services such as Azure Database for PostgreSQL and Azure Database for MySQL (flexible server).

The real decision. These two services are not like-for-like. RDS is a multi-engine, instance-based service; Azure SQL Database is a single-engine, database-scoped service. If you need SQL Server in the cloud, the closer pairings are RDS for SQL Server versus Azure SQL Managed Instance (both give you an instance), or RDS for SQL Server versus Azure SQL Database when your application fits in one database. Most teams choose the cloud first and the service second.

Side by side

Aspect	Amazon RDS	Azure SQL Database
What it is	Managed relational database instances on AWS	Managed single databases and elastic pools on Azure (PaaS)
Engines	Db2, MariaDB, SQL Server, MySQL, Oracle, PostgreSQL (Aurora is a related but separate family)	SQL Server engine only (latest stable version)
Unit you manage	A DB instance (instance class plus storage), which can hold several databases	A database or an elastic pool; no instance
SQL Server versions	You choose: SQL Server 2016, 2017, 2019, 2022 or 2025	Evergreen engine; you control behaviour with database compatibility levels
SQL Server Agent	Yes on Enterprise, Standard and Web editions, with restrictions	No; Microsoft points to Elastic jobs
Native .bak backup and restore	Yes, through Amazon S3 using RDS stored procedures	No BACKUP or RESTORE statements; export with BACPAC
High availability	Optional Multi-AZ deployment (synchronous standby, or a three-instance Multi-AZ DB cluster for MySQL and PostgreSQL)	Built into every database; zone redundancy, geo-replication and failover groups available
Automated backups	Retention 0 to 35 days (0 disables them), point-in-time restore within that period	Point-in-time restore 1 to 35 days (7 by default); long-term retention up to 10 years
Serverless compute	Not for RDS DB instances (Aurora Serverless v2 is separate)	Yes, in General Purpose and Hyperscale, with auto-pause in General Purpose
OS access option	RDS Custom for Oracle and SQL Server	None; use SQL Server on Azure VMs for OS access
Pricing model	Instance hours, provisioned storage and IOPS, backup storage, data transfer; reserved instances	vCore (provisioned or serverless) or DTU, plus storage; reservations and Azure Hybrid Benefit
Main trade-off	You size and pay for instances, and Multi-AZ is an extra cost you opt into	SQL Server only, and no instance-level features

Key differences

Engine choice: six engines versus one

RDS is the broader service. One API and console cover Db2 (Standard, Advanced and Community editions), MariaDB, SQL Server, MySQL, Oracle Database and PostgreSQL, so a team on AWS can run a PostgreSQL application and an Oracle ERP database with the same backup, monitoring and Multi-AZ model. Each engine keeps its own feature set and restrictions, documented per engine in the RDS User Guide.

Azure SQL Database runs only the SQL Server engine. On Azure, PostgreSQL and MySQL are separate products (Azure Database for PostgreSQL and Azure Database for MySQL flexible server), each with its own tiers and documentation, and there is no Azure counterpart to RDS for Oracle or RDS for Db2 in the Azure SQL family. If you need several engines on one managed platform, that points to RDS on AWS; if your estate is SQL Server, both clouds can serve it.

SQL Server on RDS versus Azure SQL Database

This is the comparison most readers actually need. RDS for SQL Server gives you a SQL Server instance: you choose the version (2016 to 2025 are listed by AWS) and edition (Express, Web, Standard and Enterprise under License Included; Enterprise, Standard and Developer under Bring Your Own Media), and you can create several databases on it (30 to 100 depending on instance class and availability mode). SQL Server Agent runs on Enterprise, Standard and Web editions, though AWS restricts it: no PowerShell or command shell job steps, no Agent email notifications, alerts or operators, and no Agent-driven backups. You can import and export native .bak files through Amazon S3.

The master user is in db_owner for user databases, not sysadmin, and AWS lists features that are not supported, including FILESTREAM, maintenance plans, log shipping, replication, Service Broker endpoints, PolyBase, server-level triggers and xp_cmdshell. CLR is not supported on RDS for SQL Server 2017 and later, linked servers have limited support, SSAS is not supported from SQL Server 2022, and SSIS is not supported on SQL Server 2025.

Azure SQL Database removes the instance altogether. Microsoft's feature comparison lists no SQL Server Agent (Elastic jobs instead), no three-part-name cross-database queries, no linked servers, CLR, Service Broker or Database Mail, and no BACKUP or RESTORE statements. In exchange you get an engine Microsoft keeps current, built-in HA for every database, serverless compute and Hyperscale databases up to 128 TB.

-- T-SQL: works on RDS for SQL Server (several databases on one instance)
-- but not in Azure SQL Database (no three-part names across databases)
SELECT o.id, c.name
FROM SalesDb.dbo.orders AS o
JOIN CrmDb.dbo.customers AS c ON c.id = o.customer_id;

-- RDS for SQL Server: native backup to S3 uses an RDS stored procedure,
-- not the BACKUP statement
EXEC msdb.dbo.rds_backup_database
     @source_db_name = 'SalesDb',
     @s3_arn_to_backup_to = 'arn:aws:s3:::my-bucket/SalesDb.bak';

In practice, RDS for SQL Server sits between Azure SQL Database and Azure SQL Managed Instance. Managed Instance supports more of SQL Server than RDS does (for example CLR, Service Broker and Database Mail), while Azure SQL Database supports less. See Azure SQL Database vs Managed Instance and Azure SQL Database vs SQL Server.

High availability and backups

On RDS, high availability is a deployment option you pay for. A Multi-AZ DB instance keeps a synchronous standby in another Availability Zone that does not serve reads; for SQL Server, AWS implements Multi-AZ with Always On availability groups or database mirroring. A Multi-AZ DB cluster (MySQL and PostgreSQL) has one writer and two readable standbys in three zones with semisynchronous replication. Read replicas are asynchronous and created manually; for SQL Server they need Enterprise edition, up to 15 per source. Automated backups use storage snapshots plus logs, with retention from 0 to 35 days (the console default is 7 days, the API default 1 day), and manual snapshots are kept until you delete them.

In Azure SQL Database, HA is part of every database. General Purpose separates compute from storage held in Azure Blob storage; Business Critical runs a cluster with up to three secondary replicas and a free read scale-out replica; Hyperscale uses a distributed architecture with optional HA replicas. Zone redundancy is available in the vCore tiers and Premium. Point-in-time restore covers 1 to 35 days (7 by default), and long-term retention keeps full backups for up to 10 years.

Maintenance and control

Every RDS DB instance has a weekly 30-minute maintenance window that you can set; required OS and database patches are applied in it, and AWS notes that mandatory upgrades can be applied after their apply date even if you keep moving the window. Multi-AZ reduces the impact of OS patching to a failover, and for SQL Server Multi-AZ, engine upgrades are rolling so the outage lasts about one failover. For MySQL, MariaDB and PostgreSQL, RDS Blue/Green Deployments can reduce upgrade downtime.

Azure SQL Database is patched continuously by Microsoft. By default, Microsoft's policy blocks most impactful updates between 8 AM and 5 PM local time; you can choose a weekday (Monday to Thursday) or weekend (Friday to Sunday) window from 10 PM to 6 AM local time, and receive advance notifications. There are no version upgrades to plan, but you also cannot hold the engine on an older version; you control behaviour through compatibility levels.

Licensing for SQL Server

RDS for SQL Server supports License Included (the SQL Server licence is in the instance price) and Bring Your Own Media for Enterprise and Standard, which requires existing licences with active Software Assurance and License Mobility; you still pay for infrastructure and the Windows licence. Developer edition is available through BYOM for non-production use. RDS Custom for SQL Server supports License Included and BYOM.

Azure SQL Database vCore prices in General Purpose and Business Critical include the SQL Server licence, and Azure Hybrid Benefit lets you apply SQL Server licences with Software Assurance for a discount. Hyperscale has no separate licence fee, so Azure Hybrid Benefit does not apply to new Hyperscale databases, and the DTU model does not support it.

Pricing and licensing

Amazon RDS. According to the AWS RDS pricing page in October 2026, you pay for DB instance hours (by instance class), provisioned storage per GB-month, Provisioned IOPS where used, backup storage for automated backups and snapshots, and data transfer. On-demand instance time is billed in one-second increments with a 10-minute minimum after a billable status change. Reserved instances for one or three years advertise savings of up to 69%, and AWS also lists Database Savings Plans. Multi-AZ, read replicas and the SQL Server or Oracle licence (under License Included) all add to the instance cost, and engines past community end of life can incur Extended Support charges. Prices vary by engine, region and instance class, so use the AWS Pricing Calculator. New AWS customers receive up to USD 200 in Free Tier credits usable on RDS, on a free plan that lasts 6 months; RDS Free Tier covers MariaDB, MySQL, PostgreSQL and SQL Server Express on micro instance classes, Single-AZ only.

Azure SQL Database. Billed per database or elastic pool. In the vCore model you pay for compute (provisioned per hour, or serverless per second of use), data storage and backup storage beyond the included amount; Microsoft Learn states prices start from USD 5 per month for the Basic (DTU) tier. Discounts come from reservations (vCore only) and Azure Hybrid Benefit (vCore, not Hyperscale). The free offer, per Microsoft Learn in October 2026, covers up to 10 General Purpose serverless databases per subscription, each with 100,000 vCore seconds, 32 GB of data and 32 GB of backup storage per month. Use the Azure pricing calculator for your region and tier.

Neither provider publishes a like-for-like comparison, and the result depends on engine, edition licensing, HA choice, region and commitment. We have not produced a worked cost comparison because neither vendor's pages support one directly.

Pricing checked on the vendors' official pages on 7 October 2026. Prices change; confirm before buying.

Where each one leads

Amazon RDS strengths

Six engines (Db2, MariaDB, SQL Server, MySQL, Oracle, PostgreSQL) under one management model
SQL Server as a real instance: SQL Server Agent, several databases per instance and native .bak import and export through S3
You choose the engine version and when to upgrade, within AWS support policies
RDS Custom gives OS and database access for Oracle and SQL Server applications that need it
Close integration with the rest of AWS (VPC, IAM, CloudWatch, S3)

Azure SQL Database strengths

High availability built into every database, with zone redundancy available in vCore tiers
Serverless compute that autoscales, with auto-pause in General Purpose
Hyperscale databases up to 128 TB
No engine upgrades to plan: Microsoft keeps the engine current
Free offer of up to 10 small serverless databases per subscription

Limitations

Amazon RDS limitations

No sysadmin and no host access on standard RDS; many SQL Server features are unsupported or limited (CLR on 2017 and later, FILESTREAM, replication, log shipping)
Multi-AZ, read replicas and licences are separate cost lines you have to plan
No serverless option for RDS DB instances (only in the separate Aurora family)
You still own engine version upgrades and must act before AWS deprecation dates

Azure SQL Database limitations

SQL Server engine only; other engines are separate Azure services
No SQL Server Agent, cross-database queries, linked servers, CLR, Service Broker or Database Mail
No native BACKUP or RESTORE, and no way to restore its backups to SQL Server directly
You cannot pin an older engine version; only compatibility levels

When to choose each

Choose Amazon RDS if

Your applications already run on AWS
You need PostgreSQL, MySQL, MariaDB, Oracle or Db2 as a managed service
You are moving a SQL Server instance that relies on SQL Server Agent jobs, several databases or .bak files
You need a specific SQL Server version or operating system level access (RDS Custom)

Choose Azure SQL Database if

Your organisation is standardised on Azure and Microsoft Entra ID
You are building a new SQL Server application that fits within one database or a pool of independent databases
Your workload is intermittent and suits serverless auto-pause, or you need very large databases with Hyperscale
You want no engine upgrades to schedule

When neither is right

You are moving a SQL Server estate to Azure with minimal changes: look at Azure SQL Managed Instance, which supports Agent, cross-database queries, CLR and native backup and restore. See Azure SQL Database vs Managed Instance.
You need full control of SQL Server, including sysadmin and the operating system: run SQL Server yourself on Amazon EC2 or on Azure VMs. See Azure SQL Database vs SQL Server.
You are on AWS and want MySQL or PostgreSQL with storage that scales automatically or serverless compute: compare Amazon RDS vs Amazon Aurora, or a serverless Postgres service in Neon vs AWS RDS.
Your organisation is on Google Cloud: see AWS RDS vs Google Cloud SQL and Azure SQL vs Google Cloud SQL.

Final recommendation

Bottom line

For most teams the cloud decides: Amazon RDS if you are on AWS, Azure SQL Database (or another Azure SQL option) if you are on Azure. Beyond that, RDS wins on engine choice and on giving you a SQL Server instance with Agent jobs and .bak files, while Azure SQL Database wins on built-in HA, serverless compute, Hyperscale and having no engine upgrades to manage. If you are comparing them for SQL Server, check your application's use of instance-level features first: if it needs them, the fair comparison is RDS for SQL Server against Azure SQL Managed Instance.

Frequently asked questions

Can I run SQL Server on Amazon RDS?

Yes. RDS for SQL Server lists SQL Server 2016, 2017, 2019, 2022 and 2025, in Express, Web, Standard and Enterprise editions under License Included, and Enterprise, Standard and Developer under Bring Your Own Media. You get an instance with SQL Server Agent (Enterprise, Standard and Web) and native .bak backup and restore through S3, but no sysadmin role or host access.

Is Azure SQL Database the same as SQL Server on RDS?

No. Both run the SQL Server engine, but Azure SQL Database is database scoped and has no SQL Server Agent, cross-database queries or native backups, while RDS for SQL Server gives you an instance with several databases and Agent jobs. Azure SQL Managed Instance is the closer Azure equivalent to an RDS SQL Server instance.

Does Azure have an equivalent of RDS for PostgreSQL and MySQL?

Yes, as separate services: Azure Database for PostgreSQL and Azure Database for MySQL (flexible server). They are not part of the Azure SQL family. Azure SQL Database itself runs only the SQL Server engine.

Which engines does Amazon RDS support?

The Amazon RDS User Guide lists IBM Db2, MariaDB, Microsoft SQL Server, MySQL, Oracle Database and PostgreSQL. The RDS product page also counts the Aurora family (Aurora MySQL-Compatible, Aurora PostgreSQL-Compatible and Aurora DSQL), which has its own user guide. RDS Custom supports only Oracle and SQL Server.

Is there a free tier for either?

AWS gives new customers up to USD 200 in Free Tier credits, on a free plan that lasts 6 months, usable on RDS (MariaDB, MySQL, PostgreSQL or SQL Server Express on micro instances, Single-AZ). Azure SQL Database has a free offer of up to 10 General Purpose serverless databases per subscription with monthly limits. Details are in the pricing section and on each vendor's page.

Sources

Checked October 2026.

How we research comparisons: our editorial method.