Bedrock Knowledge base with Normal PostgreSQL Vector DB (without Aurora)

I apologize, but I'm afraid there isn't a direct solution to use a standard PostgreSQL database (without Aurora) as a vector store for Amazon Bedrock Knowledge Bases.

Amazon Bedrock Knowledge Bases are specifically designed to work with certain vector store options, and a regular PostgreSQL database is not one of the supported options. The supported vector stores include:

1. Amazon Aurora PostgreSQL Serverless
2. Amazon OpenSearch Serverless
3. Amazon Neptune Analytics
4. Some third-party vector stores like Pinecone, Redis Enterprise Cloud, and MongoDB Atlas

For PostgreSQL-based solutions, Amazon Bedrock is configured to work with Amazon Aurora PostgreSQL, which offers additional features and optimizations for vector storage and retrieval. This is why you're seeing options for Aurora and why it's asking for a Cluster ARN.

If you want to use a PostgreSQL-based solution with Bedrock Knowledge Bases, you would need to use Amazon Aurora PostgreSQL. The quick create option for Aurora PostgreSQL sets up a serverless v2 instance, which can scale down to zero when not in use, potentially helping with costs.

If using Aurora is not an option for you, you might consider one of the other supported vector stores. However, there isn't a way to directly use a standard PostgreSQL database as a vector store for Bedrock Knowledge Bases without migrating to Aurora PostgreSQL or choosing an alternative supported vector store option.
Sources
Quick create an Aurora PostgreSQL Knowledge Base for Amazon Bedrock - Amazon Aurora
Create a knowledge base in Amazon Bedrock Knowledge Bases - Amazon Bedrock
Prerequisites for your own vector store for a knowledge base - Amazon Bedrock