Building a RAG AI Workflow with n8n, AWS ECS, Aurora PGVector Store and Amazon Bedrock
Table of Contents 1. Introduction 2. Architecture 3. Create EFS File System and update ECS Task definition 4. Setup of Aurora RDS as Vector Store 5. Create Document Bucket and Setup of SNS Topic 6. Setup of Loading Document Workflow in n8n 7. Setup of RAG Workflow in n8n 8. Conclusion 1. Introduction This article builds on a previous guide for setting up an n8n AI agent in Amazon ECS. We now present a complete solution for deploying a functional Retrieval-Augmented Generation (RAG) AI agent on AWS using n8n. As noted in the previous article, n8n workflow data is lost when ECS tasks or instances in the cluster restart. To solve this, we integrate Amazon EFS to provide persistent, shared storage for n8n metadata, ensuring data durability across restarts. We use Amazon Aurora Serverless PostgreSQL with th...