DEV Community

Gaurav Saini
Gaurav Saini

Posted on

How to build microservices with Docker - BONUS!!

Hello there šŸ‘‹

This is probably the last part in microservices basics series, but certainly not the least. Today we're going to look at how we can scale up individual services and apply load balancing. This will give a major performance boost in high traffic scenarios and increase the up-time of our application.

Without further ado, let's dive in

Spawning multiple containers

First things first, we have to create/spawn multiple containers of the service we want to apply load balancing to. Let's use the Products service for this experiment.

We can make changes to the products service definition in the docker-compose.services.yml file as follows

products:
  extends:
    file: products/docker-compose.yml
    service: app
  env_file: .env
  scale: 3    # new option "scale" to tell docker-compose to spawn 3 containers for this service
Enter fullscreen mode Exit fullscreen mode

And now, when we bring up the services using the command

docker-compose --file docker-compose.services.yml up
Enter fullscreen mode Exit fullscreen mode

we see the logs indicating there are 3 container running for the products service.

Logs showing multiple containers at startup

Awesome!! we're already halfway there.

Implementing Load Balancing

Remember when I talked about defining server groups in the nginx.conf file. Well, so far we only had one server in the group, so calling it a group didn't make much sense.

But now that we have multiple containers running, in other words, multiple servers running, we can use the server group to its fullest.

upstream products {
  server microservices-shop_products_1:4000;
  server microservices-shop_products_2:4000;
  server microservices-shop_products_3:4000;
}
Enter fullscreen mode Exit fullscreen mode

It is as simple as that! we have achieved load balancing among our 3 products service containers.

And, we have the logs to verify that

Logs demonstrating load balancing in action

These logs ā˜ļø are generated when we send multiple requests to any /products endpoint.

Since, we've only specified the names of the servers (container names and port numbers), the default algorithm used by Nginx is a Round Robin algorithm.

You can read more about the supported algorithms here.


Wanna see more

Also, if you have any suggestions, improvements, or anything more in general that you'd like to see in this series, please add a comment. I'll try my best to include it here.


Wrap up

I hope you enjoyed this series and learnt something new.

Feel free to post any questions you have in the comments below.

Cheers!

Top comments (0)