27. Dockerizing Playwright Tests

Microsoft publishes a pre-built image with browsers and all system dependencies already installed, avoiding "works on my machine" issues caused by missing OS libraries.

Pointers: The version tag (v1.48.0 here) should match your installed playwright

Python package's expected protocol version can cause obscure failures, so pin both deliberately rather than always pulling :latest.

docker run -it --rm mcr.microsoft.com/playwright/python:v1.48.0-jammy /bin/bash

What it does: Sets the base image the rest of the Dockerfile builds on top of.

Types/params: <image> (string) — image name and tag, e.g.

mcr.microsoft.com/playwright/python:v1.48.0-jammy.

Pointers: Starting from the official Playwright image (rather than a plain Python image plus manually installing browsers) is strongly preferred — it guarantees all the OS-level dependencies (fonts, codecs, etc.) that browsers need are already correctly present.

What they do:

installing pip packages)

Pointers: Copying requirements.txt and running pip install before copying the rest of the source code (as shown above) is a deliberate ordering — Docker caches each layer, so if only your test code changes (not dependencies), the dependency-install layer is reused from cache instead of re-running, meaningfully speeding up repeated builds.

• WORKDIR <path> — sets the working directory inside the container for subsequent instructions
• COPY <src> <dest> — copies files from your local machine into the image
• RUN <command> — executes a shell command during the image build (e.g.,
• CMD [...] — the default command run when a container starts from this image