There is a real push and thought process around moving as much of your boilerplate code up into your serverless cloud components as possible. By streaming DynamoDB to EventBridge Pipes, you can move a large chunk of that boilerplate into the cloud. The thinking is that for things that really don’t differentiate your solution, why not let your cloud provider take care of that for you. Their integrations are well tested, highly scalable and highly available and can be more cost effective as you don’t waste CPU cycles on things like
- Polling
- Error handling
- Data transformation
- Filtering
- Enrichment
- Event management
All of those things “could” be done say in a container or in a Lambda but again, why pay the cycles, write all of this code over and over and over when you can push it up as configuration and as a part of your CDK or SAM code that handles the deployments
As usual, if you want to skip straight to a working sample, here’s the Github repository. Feel free to pull it and then run cdk deploy npx ts-node bin/app.ts and off you go.