Docker & .Net Core WebAPI using Nginx with Loading Balancing

• This example shows you how to create a basic .Net Core WebAPI hosted in a Docker container accessed via a seperate NGINX docker container reverse proxy. There is a third docker container also using nginx to host a client test web page used to query the API.
• This is all achieved using docker-compose which references the 3 Dockerfile's to build the images and create the containers.

Building and Running

Build the docker images

$ docker-compose build

Run all three docker images

$ docker-compose up

Diagnosis

Running Containers

To check the docker containers running type the following:

$ docker ps

Logs

To check the logs get the container id from the previous command 'docker ps' and type the following:

$ docker logs --details -f -t <container id>

Terminal

To explore inside the container via a terminal type the following:

$ docker exec -it <container id>  /bin/bash

Testing

To check it all works and hit the the .net core api use the sample web client as follows:

1. Open chrome using the following command

$ chrome.exe --user-data-dir="C:/Chrome dev session" --disable-web-security

2. Go to the following in your browser:

http://localhost:13000/ 

3. Press 'Click Me'.
4. The host name of the underlying API service should be returned.

Configuration & How it works

Docker Compose - links

Inside docker-compose.yml we link the nginx proxy service to the API containers using the following:

    links:
     - api1:api1
     - api2:api2
     - api3:api3
     - api4:api4
     - api5:api5

Note: This is not required unless alias of service names are needed. However, helpful to include.

Docker Compose - expose

Inside docker-compose.yml each api uses 'expose' to make the service via the exposed port (11000) available to the linked services. Note: this does not publish the port to the host machine.

    expose:
      - "11000"

Docker Compose - ports

Inside docker-compose.yml each api the 'nginx proxy' and 'gui' services expose their internal port of 80 to the host ports 12000 & 13000 respectivly.

    ports:
      - "12000:80"

NGINX Load Balancing

Inside nginx.conf we have configured requests to 'round robin' (default when alternative not specified) between the API instances by the following configuration:

    links:
    upstream api_servers {
		server api1:11000;
		server api2:11000;
		server api3:11000;
		server api4:11000;
		server api5:11000;
    }

Then in the server section we set the following:

   proxy_pass         http://api_servers;