Building async Python applications with Tortoise ORM and Amazon Aurora DSQL

A step-by-step walkthrough shows how to build an async Python rideshare application using Tortoise ORM with Amazon Aurora DSQL. It covers the specific adaptations required for DSQL's distributed architecture: UUID primary keys instead of sequential IDs, patching asyncpg's connection pool to skip unsupported advisory locks, executing DDL statements individually since DSQL doesn't support multiple DDL statements per transaction, IAM token-based authentication via Boto3 instead of static passwords, and implementing optimistic concurrency control retry logic with exponential backoff and jitter to handle SQLSTATE 40001 serialization errors. Production considerations include connection pool tuning, token refresh strategy, observability, and DSQL's service limits (10,000 active connections, 60-minute connection cap, 3,000 row write limit per transaction).

Source: https://aws.amazon.com/blogs/database/building-async-python-applications-with-tortoise-orm-and-amazon-aurora-dsql. 8 Sync News only summarizes and links out; content copyright belongs to the authors and original sources.