When building ETL workloads in AWS, one of the first problems we usually face is not the transformation logic itself.

It is connectivity.

Before writing a complex ETL job, we need to confirm that AWS Glue can actually reach the source database, authenticate correctly, resolve the endpoint, and execute a basic query.

In this article, we will configure an AWS Glue connection to a private database, attach the proper VPC networking, configure the Glue IAM role, and create a simple validation script that can be reused before building production ETL workloads.

Requirements

AWS account

AWS Glue

A private database such as Amazon RDS, Aurora, PostgreSQL, MySQL, SQL Server, or Oracle

VPC, private subnets, and security groups

IAM role for AWS Glue

Basic knowledge of JDBC connections

The proposed architecture

The idea is simple.

AWS Glue will run inside a VPC-connected environment and use a Glue Connection to reach a private database endpoint.

The architecture will look like this:

The main goal is to…