The Rails World 2024 conference ended last week, and a lot of new components of the Rails ecosystem were released during it including Kamal 2. I decided to make some preparations for the upcoming changes. In March 2024, the Basecamp team released Thruster, a small, Go-based proxy server for Rails. While the original idea of Thruster was to provide the ability to build production-ready Rails applications from a single container for the ONCE project, it will fit perfectly into the Kamal ecosystem. Isn’t It a Redundancy? Wait? Another proxy server? Why do we need it? Allow me to explain. It doesn’t matter whether we use Traefik with Kamal v1 or Kamal-proxy with Kamal v2; inside the container, we still run Puma (or any other Ruby application server) that serves requests to the application. Besides this, Puma is also responsible for serving application assets – images, stylesheets, JavaScript files, etc. – to the clients. Before Kamal came into the scene, we usually had Nginx in our infrastructure schema. Nginx is a classic web server that can serve static files efficiently. Unfortunately, it is not adapted to container-based setups as you need to manage its configuration via a config file. However, without Nginx, Puma is not the best choice to serve static files. This is where Thruster comes into play. It wraps the Puma instance and efficiently serves static files with X-Sendfile acceleration, while Puma is responsible for the dynamic content. Thruster also provides GZip compression which was available in Traefik via middleware but won’t exist in kamal-proxy. GZip compression plays a crucial role in optimizing web performance by significantly reducing the size of files transmitted between a server and a client. By compressing assets like HTML, CSS, and JavaScript, GZIP decreases bandwidth usage and accelerates the loading times of web pages, providing users with a faster, more seamless browsing experience. This not only enhances user satisfaction but also improves search engine rankings, as faster websites tend to rank higher in search results. Additionally, reducing the size of data transfers lightens the load on servers, allowing them to handle more requests efficiently and lowering hosting costs. A reasonable question would be “Why I cannot do a GZIP compression in Rails?” Technically, you can with Rack::Deflater. But since Thruster is already responsible for serving public assets, it can also efficiently handle compression. Besides compression, Thruster also provides basic HTTP caching of public assets and HTTP/2 support. With Thruster, your Rails application can be treated as a black box, which can serve any kind of request efficiently. How to Use Thruster is straightforward to use if you’re already using Puma. First, you need to add Thruster to your Gemfile: # Gemfile

gem 'thruster' Run bundle install, and you are ready to go. Since Thruster is written in Go, the platform-specific binary will be downloaded and installed with the gem. I always create binstubs to have all the necessary commands in one place: $ bundle binstubs thruster By default, Thruster will listen on port 80, while Puma will stay on port 3000. Knowing this, we should update our application Dockerfile accordingly. I assume here that your application is already dockerized. I replaced the port that is getting exposed from 3000 to 80 and added ./bin/thrust to the default command. # Dockerfile

# Entrypoint prepares the database.
ENTRYPOINT ["/rails/bin/docker-entrypoint"]

# Start server via Thruster by default, this can be overwritten at runtime
EXPOSE 80
CMD ["./bin/thrust", "./bin/rails", "server"] The last change is to update ./bin/docker-entrypoint script. I always try to keep it as close to the original Rails template as possible. Below is the final version of the script. #!/bin/bash -e

# If running the rails server then create or migrate existing database
if [ "${@: -2:1}" == "./bin/rails" ] && [ "${@: -1:1}" == "server" ]; then
  ./bin/rails db:prepare
fi

exec "${@}" That is all! Thanks to the Traefik container-based configuration, it will read the exposed port from the container and route the traffic accordingly after you deploy the changes. After deploy is finished, you can log in to the container and check if Thruster is running: rails@docker:~$ ps aux
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
rails          1  0.0  0.3 1232848 13124 ?       Ssl  09:26   0:06 /usr/local/bundle/ruby/3.3.0/gems/thruster-0.1.8-x86_64-linux/exe/x86_64-linux/thrust ./bin/rails server Summary Since there are no visible cons of using Thruster in your Rails application, I highly recommend using it. It can improve the performance of your application and reduce the load on the server. It is also a good idea to use it as a preparation for the upcoming Kamal 2 release. The Rails World 2024 conference ended last week, and a lot of new components of the Rails ecosystem were released during it including Kamal 2. I decided to make some preparations for the upcoming changes. In March 2024, the Basecamp team released Thruster , a small, Go-based proxy server for Rails. While the original idea of Thruster was to provide the ability to build production-ready Rails applications from a single container for the ONCE project, it will fit perfectly into the Kamal ecosystem. Thruster Isn’t It a Redundancy? Wait? Another proxy server? Why do we need it? Allow me to explain. It doesn’t matter whether we use Traefik with Kamal v1 or Kamal-proxy with Kamal v2; inside the container, we still run Puma (or any other Ruby application server) that serves requests to the application. Besides this, Puma is also responsible for serving application assets – images, stylesheets, JavaScript files, etc. – to the clients. Before Kamal came into the scene, we usually had Nginx in our infrastructure schema. Nginx is a classic web server that can serve static files efficiently. Unfortunately, it is not adapted to container-based setups as you need to manage its configuration via a config file. However, without Nginx, Puma is not the best choice to serve static files. This is where Thruster comes into play. It wraps the Puma instance and efficiently serves static files with X-Sendfile acceleration, while Puma is responsible for the dynamic content. Thruster also provides GZip compression which was available in Traefik via middleware but won’t exist in kamal-proxy. GZip compression plays a crucial role in optimizing web performance by significantly reducing the size of files transmitted between a server and a client. By compressing assets like HTML, CSS, and JavaScript, GZIP decreases bandwidth usage and accelerates the loading times of web pages, providing users with a faster, more seamless browsing experience. via middleware This not only enhances user satisfaction but also improves search engine rankings, as faster websites tend to rank higher in search results. Additionally, reducing the size of data transfers lightens the load on servers, allowing them to handle more requests efficiently and lowering hosting costs. A reasonable question would be “Why I cannot do a GZIP compression in Rails?” Technically, you can with Rack::Deflater . But since Thruster is already responsible for serving public assets, it can also efficiently handle compression. Rack::Deflater Besides compression, Thruster also provides basic HTTP caching of public assets and HTTP/2 support. With Thruster, your Rails application can be treated as a black box, which can serve any kind of request efficiently. With Thruster, your Rails application can be treated as a black box, which can serve any kind of request efficiently. How to Use Thruster is straightforward to use if you’re already using Puma. First, you need to add Thruster to your Gemfile: # Gemfile

gem 'thruster' # Gemfile

gem 'thruster' Run bundle install, and you are ready to go. Since Thruster is written in Go, the platform-specific binary will be downloaded and installed with the gem. bundle install, I always create binstubs to have all the necessary commands in one place: $ bundle binstubs thruster $ bundle binstubs thruster By default, Thruster will listen on port 80, while Puma will stay on port 3000. Knowing this, we should update our application Dockerfile accordingly. I assume here that your application is already dockerized. I replaced the port that is getting exposed from 3000 to 80 and added ./bin/thrust to the default command. Dockerfile ./bin/thrust # Dockerfile

# Entrypoint prepares the database.
ENTRYPOINT ["/rails/bin/docker-entrypoint"]

# Start server via Thruster by default, this can be overwritten at runtime
EXPOSE 80
CMD ["./bin/thrust", "./bin/rails", "server"] # Dockerfile

# Entrypoint prepares the database.
ENTRYPOINT ["/rails/bin/docker-entrypoint"]

# Start server via Thruster by default, this can be overwritten at runtime
EXPOSE 80
CMD ["./bin/thrust", "./bin/rails", "server"] The last change is to update ./bin/docker-entrypoint script. I always try to keep it as close to the original Rails template as possible. ./bin/docker-entrypoint original Rails template Below is the final version of the script. #!/bin/bash -e

# If running the rails server then create or migrate existing database
if [ "${@: -2:1}" == "./bin/rails" ] && [ "${@: -1:1}" == "server" ]; then
  ./bin/rails db:prepare
fi

exec "${@}" #!/bin/bash -e

# If running the rails server then create or migrate existing database
if [ "${@: -2:1}" == "./bin/rails" ] && [ "${@: -1:1}" == "server" ]; then
  ./bin/rails db:prepare
fi

exec "${@}" That is all! Thanks to the Traefik container-based configuration, it will read the exposed port from the container and route the traffic accordingly after you deploy the changes. After deploy is finished, you can log in to the container and check if Thruster is running: rails@docker:~$ ps aux
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
rails          1  0.0  0.3 1232848 13124 ?       Ssl  09:26   0:06 /usr/local/bundle/ruby/3.3.0/gems/thruster-0.1.8-x86_64-linux/exe/x86_64-linux/thrust ./bin/rails server rails@docker:~$ ps aux
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
rails          1  0.0  0.3 1232848 13124 ?       Ssl  09:26   0:06 /usr/local/bundle/ruby/3.3.0/gems/thruster-0.1.8-x86_64-linux/exe/x86_64-linux/thrust ./bin/rails server Summary Since there are no visible cons of using Thruster in your Rails application, I highly recommend using it. It can improve the performance of your application and reduce the load on the server. It is also a good idea to use it as a preparation for the upcoming Kamal 2 release.

Walkthroughs, tutorials, guides, and tips. This story will teach you how to do something new or how to do something better.

How to Use Docker to Template Local Database

Upgrading From Rails 7.1 to Rails 7.2: A Step-by-Step Guide for Enhancing Your Application

Read About Ruby and Beyond

Too Long; Didn't Read

Make resilience your competitive advantage

Rails Thruster 101: A Quick Guide

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Untitled Story

A Beginner's Guide to Traefik Tuning for Rails Applications

The Noonification: How to Deal With Flapping or Broken Tests (11/29/2023)

The Noonification: Delving Into OpenTelemetry Collector (11/18/2023)

The Noonification: How to Implement a Merkle Tree in Solidity (11/12/2023)

105 Stories To Learn About K8s

A Beginner's Guide to Traefik Tuning for Rails Applications

The Noonification: How to Deal With Flapping or Broken Tests (11/29/2023)

The Noonification: Delving Into OpenTelemetry Collector (11/18/2023)

The Noonification: How to Implement a Merkle Tree in Solidity (11/12/2023)

105 Stories To Learn About K8s

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps