4 Dockerfile

A Dockerfile is a plain text file containing a sequence of instructions that describe, step by step, exactly how to construct a Docker image — its base, its installed dependencies, its application files, and its default runtime behavior.

The Structure of a Dockerfile

A Dockerfile is read from top to bottom, with each instruction either adding a new filesystem layer or setting metadata that will apply to containers started from the resulting image.

FROM python:3.12-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
EXPOSE 8080
CMD ["python", "app.py"]

This example starts from a base image, sets a working directory, installs dependencies, copies application source, documents the port the application uses, and defines the default startup command.

Building an Image From a Dockerfile

The docker build command reads a Dockerfile from a specified location and executes its instructions in order, producing a new image as the result.

docker build -t myapp:1.0 .

The final . argument specifies the build context — the directory whose contents are available to instructions like COPY during the build.

Why a Text File Captures the Entire Environment

Because every step needed to assemble an application's environment is written explicitly in this one file, the Dockerfile itself functions as both executable build instructions and as documentation of exactly what the image contains and how it was constructed.

cat Dockerfile

Reading a project's Dockerfile directly is often the fastest way to understand exactly what dependencies and configuration an application relies on, without needing separate documentation.

Version-Controlling the Dockerfile

Because a Dockerfile is a plain text file, it is naturally version-controlled alongside the application's source code, meaning changes to the build process are tracked with the same history and review process as any other code change.

git log -p Dockerfile

Why the Dockerfile Matters

The Dockerfile is the foundational artifact behind everything else discussed about Docker images — every layer, every piece of metadata, every base image choice traces back to instructions written in this one file, making it the primary place to look when understanding or modifying how an application is packaged.