The Amazon RDS implementation of Microsoft SQL Server on a DB instance have some limitations you should be aware of:
The maximum number of databases on a single Microsoft SQL Server DB instance is 30.
Some ports are reserved for Amazon RDS use and you can't use them when you create a DB instance.
Amazon RDS for SQL Server does not support importing data into the msdb database.
You can't rename databases on a DB instance in a SQL Server Multi-AZ with Mirroring deployment.
The_db.t1.micro_DB instance class has limited resources and is best used for testing. For example, the_db.t1.micro_DB instance class doesn't have enough resources for a full implementation of SQL Server 2012 or SQL Server 2014.
SQL Server 2016 Standard Edition only supports up to 128 GB of memory, and can be used with thedb.m4.4xlargeDB instance class.
SQL Server 2014 Standard Edition only supports up to 128 GB of memory, and can be used with thedb.m4.4xlarge, anddb.r3.4xlarge, anddb.r3.8xlargeDB instance classes.
SQL Server 2012 Standard Edition only supports up to 64 GB of memory, and can be used with thedb.m4.4xlarge, anddb.r3.2xlargeDB instance classes.
The minimum storage size for a SQL Server DB instance is 20 GB for the Express and Web Editions, and 200 GB for the Standard and Enterprise Editions.
The maximum storage size for a SQL Server DB instance is 4 TB for the Enterprise, Standard, and Web editions, and 300 GB for the Express edition.
If you have a scenario that requires a larger amount of storage, it is possible to use sharding across multiple DB instances to get around this limit. This approach requires data-dependent routing logic in applications that connect to the sharded system, so that data gets queried from and written to the appropriate shard. You can either use an existing framework likeHibernate Shardsor write custom code to enable this. If you do choose to use an existing framework, it must not require any components to be installed on the same server as the DB instance. For an example of a sharding solution using an existing framework, seeUsing an Example of Sharding with Hibernate.
Because of the extensibility limitations of striped storage attached to Windows Server, Amazon RDS doesn't currently support increasing storage on a SQL Server DB instance. We recommend that you provision storage according to anticipated future storage growth. If you need to increase the storage of a SQL Server DB Instance, you can backup your databases, create a new DB instance with increased storage, and then restore the databases into the new DB instance. For more information, seeImporting and Exporting SQL Server Databases.
Amazon RDS doesn't support some features of SQL Server. This includes components such as SQL Server Analysis Services, SQL Server Integration Services, SQL Server Reporting Services, Data Quality Services, and Master Data Services. To use these features, you can run SQL Server components in an Amazon EC2 instance with Amazon EBS storage, pursuant to Microsoft licensing policies.
Because of limitations in Microsoft SQL Server, restoring to a point in time before successful execution of a DROP DATABASE might not reflect the state of that database at that point in time. For example, the dropped database is typically restored to its state up to 5 minutes before the DROP DATABASE command was issued, which means that you can't restore the transactions made during those few minutes on your dropped database. To work around this, you can reissue the DROP DATABASE command after the restore operation is completed. Dropping a database removes the transaction logs for that database.