Using AWS Lambda with Amazon Aurora MySQL

Amazon Aurora is a fully managed relational database engine that is compatible with MySQL and PostgreSQL. In this article, I will explain how you can invoke an AWS Lambda function from an Amazon Aurora MySQL cluster with the native function lambda_sync or lambda_async .

Simple Use Case

We have an Amazon Aurora MySQL database with a table (user) for saving user information. We have a requirement that we need to notify when the application or DB users delete records from the user table.

How it works

1. We have an Aurora MySQL database with the user table.
2. There is a MySQL trigger associated with the user table.
3. This trigger will be activated when the DELETE action is executed on the table.
4. Inside the trigger, we call the lambda_sync function to invoke the Lambda function.
5. Then Lambda function sends an SNS notification.

Steps

• Create a VPC
• Create an AWS Lambda function
• Create an SNS topic
• Create an IAM role for the Aurora MySQL DB cluster
• Associate the IAM role with the Aurora MySQL DB cluster
• Create an Aurora MySQL database cluster
• Setup an EC2 instance for connecting to the database
• Create a database, table, and trigger
• Testing

Note that I use AWS CDK for creating all the resources.

Create a VPC

We deploy our Aurora MySQL DB cluster in a private subnet of a VPC. However, the Lambda function resides outside the VPC. Therefore, Aurora MySQL won’t be able to invoke the Lambda function. To fix this, we can deploy either a NAT gateway in a public subnet and set the routes or create a VPC interface endpoint between the Aurora MySQL cluster and AWS Lambda function. In this example, I deploy a NAT gateway.