I have failed numerous times over the years to get Flask, gunicorn, and NGINX working together. I will give it another go. Below is the order of how I think things should be done to keep things simple.

• Get Python and Flask installed in a container
• Get Flask working with gunicorn
• Get gunicorn working in a container
• Start implementing docker-compose
• Finally, introduce NGINX
• Get this working on AWS with a default IP
• [ ] Get NGINX working on https, port 443
• [ ] Create an AMI
• [ ] Get this working on AWS with a subdomain
• [ ] Create an EC2 with docker and docker-compose. Then you can probably launch it every time the server startes with user_data

^ It's a bit more subtle here. You want docker to be able to automatically run in two scenarios.

1. Every time a new sever is spun up based off an image. (this is the same as the TODO above - user_data is only executed when an AMI launches)
2. Anytime the server crashes and reboots itself. I first found out about this in this post. It suggests using something called systemctl

docker-compose up -d --build --scale app=3

WIP, but for now

1. Create a new EC2 from scratch (TODO: create an AMI)
2. Manually install git, docker, docker-compose and this repo (TODO: I believe an AMI also solves this)
3. Manually set the correct user permissions for the above services and start the services (TODO: I believe an AMI also solves this)
4. Set docker-compose up so that if it crashes it automatically reboots
5. Figure out the correct commands for user_data and spin up a new EC2 to test