Oatfin Under The Hood - Part I

Welcome to the 11th edition of Cloud Musings! I thought I would take a few editions to go under the hood and show what happens when a customer deploys a cloud infrastructure.

I like to think of Oatfin as a container platform or a simpler, more intuitive, and developer friendly version of Kubernetes.

Automating the deployment is not only challenging, but also time consuming and frustrating at times. In addition, it requires a lot of technical expertise. Equally challenging is the tearing down of an infrastructure. This is a massive problem when it comes to the major cloud providers (AWS, Google Cloud, and Azure) and in fact that's one of the reasons they make a lot of money.

For instance, one startup I worked at, we were spending upwards of a quarter million dollars on AWS per year, but no one knew what was going on or what servers they could shut down successfully without impacting customers.

In this edition, I'll talk about what happens behind the scenes when you deploy a cloud infrastructure with Oatfin then the next edition, I’ll write about how we un-deploy and tear down an infrastructure.

Here is what the logs look like when you deploy:

As you can see these are the steps we take in order:

1. Fetch the customers' AWS credentials from Secrets Manager

2. Create a security group for the load balancer

3. Create a security group for the API or tasks

4. Add inbound rules for the load balancer

5. Add inbound rules for the API or tasks

6. Create a load balancer

7. Create a target group

8. Add a listener to the load balancer

9. Create a logs group

10. Create ECS cluster

11. Register the new task definition

12. Create ECS service

13. Register a scalable target

14. Create an autoscaling group

15. Finally we run the task

I’ve been open sourcing some of the code as I try to de-mystify what Oatfin does under the hood. In the next edition, I will talk about the un-deploy and tearing down of the infrastructure.

Thanks for reading!

Jay