Skip to content
This repository has been archived by the owner on Sep 25, 2024. It is now read-only.

SQLAlchemy dialect to connect to Amazon RDS instances with IAM authentication. Modified to support CNAME resolution.

License

Notifications You must be signed in to change notification settings

rossumai/sqlalchemy-rdsiam

 
 

Repository files navigation

sqlalchemy-rdsiam

SQLAlchemy dialect to connect to Amazon RDS instances with IAM authentication.

The following are supported:

  • Amazon RDS PostgreSQL, with psycopg2.
  • Amazon RDS PostgreSQL, with asyncpg.

SQLAlchemy 1.3, 1.4 and 2.0 are supported.

Background

Amazon RDS is managed database service on AWS, which provides the ability to connect to database instances with IAM authentication instead of passwords. AWS SDKs or the AWS CLI can be used to generate a connection token, valid for 15 minutes, and based on an Amazon IAM identity.

With SQLAlchemy, it is possible to use IAM authentication using one of the following options:

Both options require modifying the codebase to either inject the event handler or the custom connection factory. With many open-source tools, this requires maintaining a fork. This repository provides a set of dialects which can be installed and used directly in any SQLAlchemy codebase instead.

Getting Started

  • Install the Python package:

    pip install sqlalchemy-rdsiam
  • Use a connection string with scheme corresponding to the target PostgreSQL library, and leave out the password. For instance:

    postgresql+psycopg2rdsiam://username@host/dbname
    postgresql+asyncpgrdsiam://username@host/dbname

    Note: if a password is provided, it will be ignored.

  • Run with an IAM identity that has IAM permissions to connect to the database. See IAM authentication.

Additional Configuration

AWS Region

The default region in the environment is used. To access a database in a different region without changing your environment, pass the query parameter aws_region_name in the connection string:

postgresql+psycopg2rdsiam://username@host/dbname?aws_region_name=us-east-2

Creating the Database If It Doesn't Exists

The dialect supports optionally creating the database upon connection if it doesn't exist. This is disabled by default. To create the database if it doesn't exist, set the query parameter create_db_if_not_exists to true:

postgresql+psycopg2rdsiam://username@host/dbname?create_db_if_not_exists=true

Note: the role used must have permissions to create databases.

Set sslrootcert to the Amazon RDS Certificate Bundle

Amazon RDS TLS certificates are signed by Amazon certificate authorities, and the sslrootcert PostgreSQL argument must be used in order to verify the certificate chain when connecting to the instance. In some cases, it can be useful to directly get the CA bundle along with the package for testing, or to streaming provisioning. To this end, the CA bundle is automatically downloaded when installing the Python package, and you can opt-in to use it directly.

Note: make sure this is in line with your security posture requirements first.

The package can directly set sslrootcert to the certificate bundle for all Amazon RDS regions. This is disabled by default. To do so, set the query parameter rds_sslrootcert to true:

postgresql+psycopg2rdsiam://username@host/dbname?rds_sslrootcert=true

You still need to set sslmode - for instance, with sslmode=verify-full:

postgresql+psycopg2rdsiam://username@host/dbname?rds_sslrootcert=true&sslmode=verify-full

See SSL Support for additional details.

Contributing

See Contributing.

License

See License.

About

SQLAlchemy dialect to connect to Amazon RDS instances with IAM authentication. Modified to support CNAME resolution.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%