Azure SQL Database
by Microsoft • • Relational Databases
Azure SQL Database is Microsoft's relational cloud database service for Transact-SQL apps, run as a platform service with vCore, serverless and DTU options.
Azure SQL Database gives an application a relational database in Microsoft's cloud without a server to run. A team writes Transact-SQL, and Azure handles high availability, automated backups and failover. Microsoft builds and operates the service within the Azure SQL family, beside Azure SQL Managed Instance and SQL Server on Azure Virtual Machines, which are separate offerings. A database is created through the Azure portal, a PowerShell script or the Azure CLI. Each database is deployed on its own or inside an elastic pool, where databases on one server share a set of resources. Compute comes in the General Purpose, Business Critical and Hyperscale tiers, plus a serverless option and the DTU model of fixed compute bundles. Hyperscale separates storage from compute and offloads reads to named replicas. Active geo-replication keeps readable secondaries in other regions, and after a failover, connections are redirected to the new primary automatically. Inside the engine sit Query Store, intelligent query processing, columnstore indexes, a native JSON type, full-text search and vector search. Transparent data encryption, row-level security and Microsoft Defender for SQL protect the data. The service suits developers and database teams whose applications already speak Transact-SQL and who want the database run for them. Software companies with many tenant databases can group them in elastic pools. Hyperscale fits very large or fast-growing databases, while serverless fits databases that sit idle between bursts of use. Teams already running workloads on Azure get it inside the same account and portal they use for everything else. Typical work is the data tier behind a cloud-native web app, or a SQL Server workload moved off its own servers. Mirroring in Fabric copies data in near real time into Microsoft Fabric, where analysts query it. Developers connect it to Azure Functions without writing their own code, and can turn database objects into REST and GraphQL APIs. For AI features, native vector search works with Azure OpenAI and Azure AI Search on the application's own SQL data.
Pricing
View plans-
Free offer
Free
- Up to 10 General Purpose serverless databases per Azure subscription
- 100,000 vCore seconds of compute, 32 GB of data and 32 GB of backup storage per database each month
- Free for the life of the subscription
- New Azure accounts also get a $200 credit to use within 30 days
-
General Purpose provisioned (vCore)
Custom pricing
- $0.30443 an hour of compute for 2 vCores, Azure retail price for East US, with the SQL license billed per vCore-hour on top
- Compute only; storage and backup storage billed per GB
- Sizes up to 80 vCores, with remote-storage high availability
-
General Purpose serverless (vCore)
Custom pricing
- $0.52176 per vCore-hour, billed per second of use, Azure retail price for East US
- Zone redundancy adds $0.31306 per vCore-hour
- Single databases only; compute scales up and down with use
-
Business Critical provisioned (vCore)
Custom pricing
- $0.60887 an hour of compute for 2 vCores, Azure retail price for East US, with the SQL license billed per vCore-hour on top
- Local-storage high availability with readable secondary replicas
- Sizes up to 80 vCores; compute only, storage billed separately
Features
- Included: ACID transactions with row-level locking
- Included: Declarative table partitioning
- Included: Native JSON column type with indexing
- Included: Built-in full-text search indexes
- Included: Streaming physical replication to standbys
- Not included: Logical replication of selected tables
- Included: Synchronous multi-replica commit
- Included: Automatic failover to a standby
- Included: Point-in-time recovery from archived logs
- Included: Online schema changes without table locks
- Included: Columnstore indexes for analytics
- Included: Row-level security policies
- Included: Transparent data encryption at rest
- Not included: Stored procedures in multiple languages
- Not included: Extension or plugin architecture
- Included: Query store or plan history
- Not included: In-place major version upgrade tooling
- Not included: Open source core under a permissive license
Additional Features
- Single databases and elastic pools
- General Purpose, Business Critical and Hyperscale service tiers
- Serverless compute that scales with use
- DTU purchasing model
- Hyperscale storage up to 100 TB
- Hyperscale named replicas for read scale-out
- Active geo-replication with readable secondaries
- Failover groups
Best for
- Applications written in Transact-SQL moving to a managed cloud service
- Teams already running their workloads on Azure
- SaaS products running many tenant databases in elastic pools
- Databases used in bursts that suit serverless compute
- Very large or fast-growing databases on Hyperscale
- AI applications that keep vectors beside relational data
- Analytics teams mirroring operational data into Microsoft Fabric
Use cases
- Hosting the database behind a cloud-native web app
- Moving a SQL Server database to a managed cloud service
- Running a multi-tenant SaaS data tier in elastic pools
- Offloading read traffic to Hyperscale named replicas
- Keeping a readable copy in a second region for disaster recovery
- Restoring a database to a point in time after a bad change
- Feeding near real-time data into Microsoft Fabric for analysis
- Building retrieval for AI features with vector search
- Exposing tables as REST or GraphQL APIs
- Enforcing per-tenant row access with row-level security
- Finding and pinning regressed query plans with Query Store
- Creating databases from the Azure portal, PowerShell or the Azure CLI
Screenshots & Videos
Explore Azure SQL Database in action
User Reviews
Compare with Similar Software
Click to compare side by side with Azure SQL Database
