What will change when using AWS API Gateway and connecting an existing API endpoint?

Question

Amazon Web Services

Hello everyone,

let's say I have my own Restful API endpoint available at https://api.mysimpleserver.ru/orders/vasya, which is tied to an EC2 instance with elastic IP and a domain (NGINX/Django REST framework application). SSL termination on NGINX web server.

In general terms, how to make this endpoint work through the API Gateway and be scalable and available? Probably it is better to use Application Load Balancer and several instances in different AZ?

I don’t quite understand what architecture can be in this case! What to register in API Gateway? Will the appearance of this endpoint change?

beaver to all)
Vadim

Answer 1

The GW API will not give scalability because you have 1 virtual machine. Your options are:
- ALB + Autoscaling Group + EC2 AMI
- ALB + Autoscaling Group + ECS ( + Fargate )
- API GW + Lambda
API GW is essentially a pumped reverse proxy and has nothing to do with scalability. Its main functionality is traffic roaming configuration in a convenient way + security + data transformation.

Answer 2

The cheapest scaling option is to rewrite the vasya REST application on AWS Lambda. This is not always architecturally possible due to session and resource limitations. But if it works out, then you can run up to 1000 copies of Vasya in parallel.