Locust base Docker 镜像如何添加自定义诗歌环境？

Question

How do you run Locust (load testing tool) in a stable Docker container with extra poetry dependencies installed? From the docs it's known that running Locust in Docker is easily possible through their base image.

docker run -p 8089:8089 -v $PWD:/mnt/locust locustio/locust -f /mnt/locust/locustfile.py

But if a load testing Python project requires extra libraries which are managed through poetry, the locust command must be run through poetry run locust . The locust docs only give the following example, but with pip :

FROM locustio/locust
RUN pip3 install some-python-package

It get's more tricky if you want to bind mount a directory to the container, as Poetry environments are linked to the working directory they're created in.

Answer 1

Assuming a python project like this that uses poetry:

.
├── Dockerfile
├── README.md
├── data
├── poetry.lock
├── pyproject.toml
├── reports (mounting this directory to Docker to save reports)
└── src (contains the locust source code)
    ├── auth.py
    ├── config.py
    ├── locustfile.py
    ├── shapes
    ├── tasks
    └── users.py

SETUP

The contents of my Dockerfile are the following:

FROM locustio/locust

USER $USERNAME

# Use non-root user as a security measure
RUN groupadd -g 61000 docker
RUN useradd -g 61000 -l -m -s /bin/false -u 61000 docker

# Setup and install poetry
RUN apt update && \
    apt install -y git curl && \
    pip install poetry

# cacheing project requirements
WORKDIR /code
COPY poetry.lock pyproject.toml /code/

# installing project requirements
RUN poetry install --no-interaction --no-ansi --no-root

COPY . /code

ENTRYPOINT [ "poetry", "run", "locust" ]

Notes:

• USER $USERNAME is important to get the correct permissions
• The poetry environment is linked to the code/ directory and it's also where the entire project's code is moved to
• The entry point ENTRYPOINT [ "poetry", "run", "locust" ] is a good pattern as extra arguments can be passed in through the docker run command.

Running the Container

Build the image with:

docker build -t locust .

Example command to run the load test:

docker run -p 8089:8089 --mount type=bind,src=$PWD/reports,dst=/code/reports locust -f src/locustfile.py --headless --host <HOST> --csv=reports/<PREFIX>

Hope that helps!