Which solution will meet these requirements with minimal effort?
Create an Amazon CloudWatch Events rule with the operations that need to be tracked on Amazon RDS. Create an AWS Lambda function to act on these rules and write the output to the tracking systems.
Create an AWS Lambda function to trigger on AWS CloudTrail API calls. Filter on specific RDS API calls and write the output to the tracking systems.
Create RDS event subscriptions. Have the tracking systems subscribe to specific RDS event system notifications.
Write RDS logs to Amazon Kinesis Data Firehose. Create an AWS Lambda function to act on these rules and write the output to the tracking systems.
Explanations:
While creating a CloudWatch Events rule can track certain events, it may not capture all database operations (like shutdown or deletion) comprehensively, and it requires setting up a Lambda function which adds additional complexity.
Using AWS CloudTrail can track API calls, but it may not provide real-time notifications for events like database shutdowns or backups. Also, filtering API calls and writing output involves more setup and isn’t as straightforward as using event subscriptions.
RDS event subscriptions provide a built-in mechanism to track specific database events such as creation, deletion, and backup. This solution is designed for tracking RDS events and can directly notify tracking systems with minimal effort.
Writing RDS logs to Kinesis Data Firehose and using a Lambda function to process them introduces unnecessary complexity. It requires additional setup and doesn’t provide a direct way to subscribe to specific RDS events.
In my experience, the answer is:
Create RDS event subscriptions. Have the tracking systems subscribe to specific RDS event system notifications.