Skip to content

Latest commit

 

History

History
205 lines (144 loc) · 7.02 KB

README.md

File metadata and controls

205 lines (144 loc) · 7.02 KB
Raw

DockerShrink

Talk to us on Slack | Vote for Feature Requests

Dockershrink is an AI-powered Commandline Tool that helps you reduce the size of your Docker images

Typical interaction with dockershrink CLI

It combines the power of traditional Rule-based analysis with Generative AI to apply state-of-the-art optimizations to your Image configurations 🧠

Dockershrink can automatically apply techniques like Multi-Stage builds, switching to Lighter base images like alpine and running dependency checks. PLUS a lot more is on the roadmap 🚀

Currently, the tool only supports NodeJS applications.

Important

Dockershrink is BETA software.

You can provide your feedback by creating an Issue in this repository.

Why does dockershrink exist?

Every org using containers in development or production environments understands the pain of managing hundreds or even thousands of bloated Docker images in their infrastructure.

High data storage and transfer costs, long build times, underprodctive developers - we've seen it all.

The issue becomes even more painful and costly with interpreted languages such as Nodejs & Python. Apps written in these languages need to pack the interpreters and all their dependencies inside their container images, significantly increasing their size.

But not everyone realizes that by just implementing some basic techniques, they can reduce the size of a 1GB Docker image down to as little as 100 MB!

(I also made a video on how to do this.)

Imagine the costs saved in storage & data transfer, decrease in build times AND the productivity gains for developers 🤯

Dockershrink aims to automatically apply advanced optimization techniques so engineers don't have to waste time on it and the organization still saves 💰!

You're welcome 😉

How it works

When you invoke the dockershrink CLI on your project, it analyzes code files.

Dockershrink looks for the following files:

👉 Dockerfile (Required)

👉 package.json (Optional)

👉 .dockerignore (Optional, created if it doesn't already exist)

It then creates a new directory (default: dockershrink.optimized) inside the project, which contains modified versions of your configuration files that will result in a smaller Docker Image.

The CLI outputs a list of actions it took over your files.

It may also include suggestions on further improvements you could make.

Installation

You can install dockershrink using PIP or PIPX

$ pip install dockershrink
$ pipx install dockershrink

Alternatively, you can also install it using Homebrew:

brew install duaraghav8/tap/dockershrink

But you should prefer to use pip instead because installation via brew takes a lot longer and occupies significanlty more space on your system (see this issue)

Usage

Navigate into the root directory of one of your Node.js projects and invoke dockershrink with the optimize command:

$ dockershrink optimize

Dockershrink will create a new directory with the optimized files and output the actions taken and (maybe) some more suggestions.

For detailed information about the optimize command, run

dockershrink optimize --help

You can also use the --verbose option to get stack traces in case of failures:

$ dockershrink optimize --verbose

To enable DEBUG logs, you can set the environment variable

export DOCKERSHRINK_CLI_LOGLEVEL=DEBUG
dockershrink optimize

Using AI Features

Note

Using AI features is optional, but highly recommended for more customized and powerful optimizations.

To use AI, you need to supply your own OpenAI API key, so even though Dockershrink itself is free, openai usage might incur some cost for you.

By default, dockershrink only runs rule-based analysis to optimize your image definition.

If you want to enable AI, you must supply your OpenAI API Key.

dockershrink optimize --openai-api-key <your openai api key>

# Alternatively, you can supply the key as an environment variable
export OPENAI_API_KEY=<your openai api key>
dockershrink optimize

Note

Dockershrink does not store your OpenAI API Key.

So you must provide your key every time you want "optimize" to use AI features. This is to avoid any unexpected costs.

Default file paths

By default, the CLI looks for the files to optimize in the current directory.

You can also specify the paths to all files using options (see dockershrink optimize --help for the available options).

Development 💻

Note

This section is for authors and contributors. If you're simply interested in using Dockershrink, you can skip this section.

  1. Clone this repository
  2. Navigate inside the root directory of the project and create a new virtual environment
python3 -m venv .venv
source .venv/bin/activate
  1. Install all dependencies
pip install --no-cache-dir -r requirements.txt
  1. Install the editable CLI tool
# -e ensures that the tool is editable, ie, code changes reflect in the tool immediately, without having to re-install it
pip install -e .

# Try running the cli
dockershrink --help
  1. Make your code changes
  2. Run black
black .
  1. In case of any changes in dependencies, update requirements.txt
pip freeze > requirements.txt

Release 🚀

Once all code changes have been made for the next release:

Then proceed to follow these steps to release new dockershrink version on PyPI:

  1. Build dockershrink from source
python -m build
  1. Upload to testpypi
twine upload --repository testpypi dist/*
  1. The new version of the package should now be available in TestPyPI.
  2. Try installing the test package
pip install --index-url https://test.pypi.org/simple/ --no-deps dockershrink
  1. Create a new Tag and push it
git tag -a <VERSION> -m "Tag version <VERSION>"
git push origin <VERSION>
  1. Create a new release in this repository
  2. Upload the package to PyPI
twine upload dist/*
  1. The new version of the package should now be available in PyPI
  2. Update the package in Dockershrink Homebrew Tap as well.