Building agentic AI patterns with Amazon Bedrock and SQL Server 2025 on Amazon RDS

This article demonstrates how SQL Server 2025 on Amazon RDS can invoke Amazon Bedrock foundation models directly from T-SQL using the new sp_invoke_external_rest_endpoint system stored procedure, eliminating the need for middleware integration layers.

• SQL Server 2025 enables native HTTPS REST API calls directly from T-SQL stored procedures without Lambda or API Gateway
• Database Scoped Credentials encrypt API keys at rest, preventing exposure in query text, plan cache, or logs
• Use cases include real-time support ticket triage, automated performance diagnostics, intelligent alerts, and on-demand query advisors
• Implementation requires enabling external REST endpoint feature via RDS parameter group and creating secure credentials
• Production-ready stored procedures demonstrate AI-powered query advisor, wait stats analysis, ticket classification, and event-driven alerts
• Default 30-second timeout is sufficient for Bedrock responses; consider SQS queuing for high-throughput scenarios

This approach brings AI capabilities directly into database workflows, reducing latency and operational complexity while maintaining security through encrypted credential management.