# Compose Sample Application ## NGINX Reverse Proxy -> WSGI -> Python/Flask Backend Project structure: ```text . ├── docker-compose.yaml ├── flask │   ├── app.py │   ├── Dockerfile │   ├── requirements.txt │   └── wsgi.py └── nginx ├── default.conf ├── Dockerfile ├── nginx.conf    └── start.sh ``` [_docker-compose.yaml_](docker-compose.yaml) ```yml services: nginx-proxy: build: nginx ports: - 80:80 flask-app: build: flask ... ``` The compose file defines an application with two services `nginx-proxy` and `flask-app`. When deploying the application, docker-compose maps port 80 of the web service container to port 80 of the host as specified in the file. Make sure port 80 on the host is not being used by another container, otherwise the port should be changed. ## Deploy with docker-compose ```bash $ docker-compose up -d Creating network "nginx-wsgi-flask_default" with the default driver Building flask-app ... Building nginx-proxy ... Creating nginx-wsgi-flask_flask-app_1 ... done Creating nginx-wsgi-flask_nginx-proxy_1 ... done ``` ## Expected result Listing containers must show two containers running and the port mapping as below: ```bash $ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES bde3f29cf571 ...nginx-proxy "/docker-entrypoint.…" About a minute ago Up About a minute (healthy) 0.0.0.0:80->80/tcp ...nginx-proxy_1 86c44470b547 ...flask-app "gunicorn -w 3 -t 60…" About a minute ago Up About a minute (healthy) 5000/tcp, 0.0.0.0:8000->8000/tcp ...flask-app_1 ``` After the application starts, navigate to `http://localhost:80` in your web browser or run: ```bash $ curl localhost:80 Hello World! ``` Stop and remove the containers ```bash $ docker-compose down Stopping nginx-wsgi-flask_nginx-proxy_1 ... done Stopping nginx-wsgi-flask_flask-app_1 ... done Removing nginx-wsgi-flask_nginx-proxy_1 ... done Removing nginx-wsgi-flask_flask-app_1 ... done Removing network nginx-wsgi-flask_default ``` ## About By following the steps above, you will have an NGINX Reverse Proxy and a Flask backend. The general traffic flow will look like the following: `Client -> NGINX -> WSGI -> Flask` ### NGINX With this deployment model, we use NGINX to proxy and handle all requests to our Flask backend. This is a powerful deployment model as we can use NGINX to cache responses or even act as an application load balancer between multiple Flask backends. You could also integrate a Web Application Firewall into NGINX to protect your Flask backend from attacks. ### WSGI WSGI (Web Server Gateway Interface) is the interface that sits in between our NGINX proxy and Flask backend. It is used to handle requests and interface with our backend. WSGI allows you to handle thousands of requests at a time and is highly scalable. In this `docker-compose` sample, we use Gunicorn for our WSGI. ### Flask Flask is a web development framework written in Python. It is the "backend" which processes requests. A couple of sample endpoints are provided in this `docker-compose` example: * `/` - Returns a "Hello World!" string. * `/cache-me` - Returns a string which is cached by the NGINX reverse proxy. This demonstrates an intermediary cache implementation. * `/info` - Returns informational headers about the request. Some are passed from NGINX for added client visibility.