Mynth Base Container Images

This repository holds the base container images used by Mynth. These images are publicly hosted on Quay and can be accessed using the quay.io/mynth/<name> registry.

Node

The node images are lightweight and optimized containers for running node.js applications. Two tags exist for the node container:

• quay.io/mynth/node:22-base
• quay.io/mynth/node:22-dev

Usage

To use a node image, create a Dockerfile in your project directory that takes advantage of Docker’s multi-stage feature. The first stage builds your application, and the second stage copies the built files for deployment. This results in a lightweight container image.

The first part of the container uses the dev tag to build the application:

FROM quay.io/mynth/node:22-dev as builder

WORKDIR /app
COPY --chown=noddy:noddy package*.json ./
RUN npm ci

First, copy your package.json and package-lock.json to the /app directory, then run npm ci. Next, build the final production version of the application:

COPY --chown=noddy:noddy . ./
RUN npx next build && npm ci --omit dev

Run your application’s build process and uninstall developer tools with npm ci --omit dev or yarn install --production.

Now that your application is built, copy the built files to the image with the base tag:

FROM quay.io/mynth/node:22-base
WORKDIR /app
COPY --from=builder --chown=noddy:noddy /app ./

Copy the files from the /app directory in your builder container, as well as all the source code files from your local repository.

Your application is now ready to run, so include a command and expose any necessary ports:

EXPOSE 3000
CMD ["next", "start"]

Build the Dockerfile as usual:

docker build -t node-example .

Now you can run your application:

docker run -p 3000:3000 node-example

If you follow the example provided in examples/node, you can access the running web application at http://localhost:3000/.

Python

The python image is a lightweight and optimized container for running Python applications that use poetry. It comes with Python 3.11 installed.

Usage

To use the python image, create a Dockerfile in your project directory that utilizes the multi-stage feature of Docker. The first stage will build your application, and the second stage will copy the built files for deployment. This results in a lightweight container image.

The first part of the container will use the dev tag to build the application:

FROM quay.io/mynth/python:dev as builder

COPY poetry.lock pyproject.toml /app/
RUN install-poetry-app hello_python

The install-poetry-app script helps install your application. First, copy your poetry.lock and pyproject.toml to the /app/ directory, then call install-poetry-app with the name of your application.

Now that your application is built, you can copy the built files to the image with the base tag:

FROM quay.io/mynth/python:base

COPY --from=builder /app /app
COPY hello_python /app/hello_python

Copy the files from the /app directory in your builder container, as well as all the source code files from your local repository.

Now your application is ready to run, so you can include a command and expose any necessary ports:

EXPOSE 8000
CMD ["uvicorn", "--host", "0.0.0.0", "hello_python.app:app"]

Build the Dockerfile as usual:

docker build -t python-example .

Now your application can be run:

docker run -p 8000:8000 python-example

If you follow the example provided in examples/python, you’ll be able to access the running web application at http://localhost:8000/.

Embracing Ubuntu as the Ideal Base for Container Images

In the world of containerization, choosing the right base image is essential for achieving a balance between security, productivity, and developer friendliness. Our team has selected Ubuntu as the base container image. We will briefly discuss the reasons behind this decision, comparing it to other popular alternatives such as Alpine, Debian, and Distroless.

The Case for Ubuntu

Security and Stability

Ubuntu is a widely-used and well-tested distribution, ensuring that applications have been proven in real-world scenarios. This minimizes the chances of encountering rare bugs, which can be a concern with less mainstream distributions like Alpine. Additionally, Ubuntu is known for its quick updates, including security patches, which is crucial for maintaining a secure environment.

Improved Productivity

Compared to Alpine, which can result in longer build times and introduce bugs due to its use of musl, Ubuntu offers extensive pre-existing library support, leading to faster build times. This increased productivity is a significant advantage for developers working with container images.

Familiarity and Long-term Use

Our team has been using Ubuntu as a base operating system for servers for many years. This familiarity with the distribution and its ecosystem allows us to leverage our existing knowledge and expertise, further enhancing productivity and efficiency. Additionally, many other users worldwide use Ubuntu, making it easier to find resources on the subject to help debug problems.

Comparing Alternatives

Alpine

While Alpine reduces the attack vector and can result in minimal-sized images, it can also hinder productivity with longer build times and introduce bugs due to its use of musl. This trade-off makes Alpine a less attractive option compared to Ubuntu.

Debian

Debian releases new versions more slowly than Ubuntu, resulting in older packages and software versions. This can sometimes lead to bugs or security vulnerabilities. Moreover, Ubuntu’s rapid update cycle ensures that security vulnerabilities are addressed promptly, whereas Debian may take longer to patch its packages.

Distroless

Distroless images, based on Debian, share many of the same advantages as Debian-based container images like Ubuntu. However, they can reduce productivity due to the removal of many useful tools from the container images. While this may provide a small security advantage, it can also hinder developers when debugging issues. The increased development overhead and minimal security benefits make Distroless a less appealing choice compared to Ubuntu.