[Misc] Reduced runtime's container image size

[nwangfw]

Pull Request Description

Enabled multi-stage docker builds and the runtime image is reduced to 1.13 GB.
Github Action Installation-test is also updated to cover the Dockerfile changes.

Related Issues

Resolves: #496

Important: Before submitting, please complete the description above and review the checklist below.

---

Contribution Guidelines (Expand for Details)

We appreciate your contribution to aibrix! To ensure a smooth review process and maintain high code quality, please adhere to the following guidelines:

Pull Request Title Format

Your PR title should start with one of these prefixes to indicate the nature of the change:

• [Bug]: Corrections to existing functionality
• [CI]: Changes to build process or CI pipeline
• [Docs]: Updates or additions to documentation
• [API]: Modifications to aibrix's API or interface
• [CLI]: Changes or additions to the Command Line Interface
• [Misc]: For changes not covered above (use sparingly)

Note: For changes spanning multiple categories, use multiple prefixes in order of importance.

Submission Checklist

• PR title includes appropriate prefix(es)
• Changes are clearly explained in the PR description
• New and existing tests pass successfully
• Code adheres to project style and best practices
• Documentation updated to reflect changes (if applicable)
• Thorough testing completed, no regressions introduced

By submitting this PR, you confirm that you've read these guidelines and your changes align with the project's contribution standards.

[Jeffwan]

I feel this is kind of hacky. Can we build the aibrix pythong package and copy that one and install? Instead of copying all the python path, I feel that way is much more clean and reliable

[nwangfw]

Great Suggesetions, Thanks! The new image built in the below way is a little bit larger. Let me know if you have any way to further optimize the below Dockerfile.

ARG PYTHON_VERSION=3.11
ARG BASE_IMAGE=python:${PYTHON_VERSION}-slim-bookworm

# Builder stage
FROM ${BASE_IMAGE} AS builder

WORKDIR /app

ARG POETRY_VERSION=1.8.3

# Install dependencies
RUN apt-get update \
    && apt-get install -y python3-dev build-essential \
    && apt-get clean && \
    rm -rf /var/lib/apt/lists/*
    
# Install Poetry
RUN python3 -m pip install poetry==${POETRY_VERSION}

# Copy the runtime source
COPY python/aibrix/poetry.lock python/aibrix/pyproject.toml python/aibrix/ /app/

# Install dependencies and build package
RUN poetry config virtualenvs.create false \
    && poetry install --no-interaction --no-dev \
    && poetry build -f wheel

# Final stage
FROM ${BASE_IMAGE}

WORKDIR /app

# Install build dependencies temporarily
RUN apt-get update \
    && apt-get install -y --no-install-recommends gcc python3-dev \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

# Copy only the built wheel and install with dependencies
COPY --from=builder /app/dist/*.whl ./
RUN pip install --no-cache-dir ./*.whl \
    && rm -f ./*.whl \
    && apt-get purge -y gcc python3-dev \
    && apt-get autoremove -y

# Set entrypoint for Runtime
ENTRYPOINT ["aibrix_runtime"]

Here is a comparison of three different approaches:
Current image size (2.53 GB), image size built by this new method (1.74 GB), image size built by the hacky method (1.13 GB)

Which way do you think is the best now? @Jeffwan

[Jeffwan]

Let's have a short discussion in office @nwangfw

[Jeffwan]

The tricky part is the docker images is build layer by layer. you have three parts separately. Even you install & remove the gcc, python-dev later but it still take one layer instead. We either need to use 3rd party tools to squash the layers or put them in the same layer.

I combine commands into 1 and get results below

Even it's still a little bit larger than PYTHONPATH version, I think it's acceptable. (Technically, we can make it close by removing all cache files but that's change Dockerfile a lot which is not necessary at this moment)

[nwangfw]

Thanks for helping me figure it out! The modified Dockerfile looks very close to PYTHONPATH version!

[Jeffwan]

/lgtm