Hello everyone 👋 I’m working on a proof of concept for Prefect using AWS and am having a little bit of conceptual difficulty understanding how all of the prefect pieces interact with each other. Additionally, I am curious what development workflows people use to register additional flows (git or ECR). Here is what I’m currently trying and want to validate that this is the prefect way to do it. I want to put prefect server on ec2 (until we are approved for cloud), run an agent (or a few) on separate ec2 instances. Register flows to server using the aws cli (that contain fargate environment specifics). Am I on the right track or should I try something different?

Hi @Jacob (he/him), welcome! I'll let the folks in our community chime in on how they manage their development workflows, but it sounds like you're on the right track. As long as you're able to communicate with the Server API on your EC2 instance, you should be in good shape. 👍 It sounds like you're aiming for a setup where Server and your agent run on EC2 instances, with actual flow deployment executing on Fargate, correct?

That’s right. It seems like a lot of the utility of prefect is being able to dynamically allocate resources at the flow level, which is a good use case for fargate.

Agreed! Then as long as your Prefect Server endpoint is accessible to the Fargate agent, you should be good to go.