pr-sglang.yml

name: PR - SGLang

on:
  pull_request:
    branches: [main]
    types: [opened, reopened, synchronize]
    paths:
      - "**sglang**"

permissions:
  contents: read
  pull-requests: read

env:
  # CI Image configuration
  CONTAINER_TYPE: "general"
  FRAMEWORK: "sglang"
  SGLANG_VERSION: "0.5.5"
  PYTHON_VERSION: "py312"
  CUDA_VERSION: "cu129"
  OS_VERSION: "ubuntu22.04"
  # Prod Image configuration
  PROD_SAGEMAKER_IMAGE: sglang:0.5-gpu-py312
  # CI environment configuration
  FORCE_COLOR: "1"
  TEST_ARTIFACTS_DIRECTORY: "/test_artifacts/sglang"

jobs:
  gatekeeper:
    runs-on: ubuntu-latest
    concurrency:
      group: ${{ github.workflow }}-gate-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    steps:
      - name: Checkout base branch (safe)
        uses: actions/checkout@v5
        with:
          ref: ${{ github.event.pull_request.base.sha }}
          fetch-depth: 1

      - name: Run permission gate (from base)
        uses: ./.github/actions/pr-permission-gate

  check-changes:
    needs: [gatekeeper]
    if: success()
    runs-on: ubuntu-latest
    concurrency:
      group: ${{ github.workflow }}-check-changes-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    outputs:
      build-change: ${{ steps.changes.outputs.build-change }}
      test-change: ${{ steps.changes.outputs.test-change }}
    steps:
      - name: Checkout DLC source
        uses: actions/checkout@v5

      - name: Setup python
        uses: actions/setup-python@v6
        with:
          python-version: "3.12"

      - name: Run pre-commit
        uses: pre-commit/action@v3.0.1
        with:
          extra_args: --all-files

      - name: Detect file changes
        id: changes
        uses: dorny/paths-filter@v3
        with:
          filters: |
            build-change:
              - "docker/sglang/**"
              - "scripts/sglang/**"
              - "scripts/common/**"
              - "scripts/telemetry/**"
              - ".github/workflows/pr-sglang*"
            test-change:
              - "test/sglang/**"

  # ======================================================
  # =============== SGLang SageMaker jobs ================
  # ======================================================
  build-sglang-image:
    needs: [check-changes]
    if: needs.check-changes.outputs.build-change == 'true'
    runs-on:
      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
        fleet:x86-build-runner
        buildspec-override:true
    concurrency:
      group: ${{ github.workflow }}-build-sglang-image-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    outputs:
      ci-image: ${{ steps.image-uri-build.outputs.CI_IMAGE_URI }}
    steps:
      - uses: actions/checkout@v5
      - run: .github/scripts/buildkitd.sh

      - name: ECR login
        uses: ./.github/actions/ecr-authenticate
        with:
          aws-account-id: ${{ vars.CI_AWS_ACCOUNT_ID }}
          aws-region: ${{ vars.AWS_REGION }}

      - name: Resolve image URI for build
        id: image-uri-build
        run: |
          CI_IMAGE_URI=${{ vars.CI_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_REGION }}.amazonaws.com/ci:sglang-${{ env.SGLANG_VERSION }}-gpu-${{ env.PYTHON_VERSION }}-${{ env.CUDA_VERSION }}-${{ env.OS_VERSION }}-sagemaker-pr-${{ github.event.pull_request.number }}
          echo "Image URI to build: ${CI_IMAGE_URI}"
          echo "CI_IMAGE_URI=${CI_IMAGE_URI}" >> ${GITHUB_ENV}
          echo "CI_IMAGE_URI=${CI_IMAGE_URI}" >> ${GITHUB_OUTPUT}

      - name: Build image
        run: |
          # base image: https://hub.docker.com/r/lmsysorg/sglang/tags
          docker buildx build --progress plain \
            --build-arg CACHE_REFRESH="$(date +"%Y-%m-%d")" \
            --build-arg BASE_IMAGE="lmsysorg/sglang:v${{ env.SGLANG_VERSION }}-${{ env.CUDA_VERSION }}-amd64" \
            --build-arg CONTAINER_TYPE="${{ env.CONTAINER_TYPE }}" \
            --build-arg FRAMEWORK="${{ env.FRAMEWORK }}" \
            --build-arg FRAMEWORK_VERSION="${{ env.SGLANG_VERSION }}" \
            --cache-to=type=inline \
            --cache-from=type=registry,ref=${CI_IMAGE_URI} \
            --tag ${CI_IMAGE_URI} \
            --target sglang-sagemaker \
            -f docker/sglang/Dockerfile .

      - name: Container push
        run: |
          docker push ${CI_IMAGE_URI}
          docker rmi ${CI_IMAGE_URI}

  set-test-environment:
    needs: [check-changes, build-sglang-image]
    if: |
      always() && !failure() && !cancelled() &&
      (needs.check-changes.outputs.build-change == 'true' || needs.check-changes.outputs.test-change == 'true')
    runs-on: ubuntu-latest
    concurrency:
      group: ${{ github.workflow }}-set-test-environment-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    outputs:
      aws-account-id: ${{ steps.set-env.outputs.AWS_ACCOUNT_ID }}
      image-uri: ${{ steps.set-env.outputs.IMAGE_URI }}
    steps:
      - name: Checkout code
        uses: actions/checkout@v4

      - name: Set test environment
        id: set-env
        run: |
          if [[ "${{ needs.build-sglang-image.result }}" == "success" ]]; then
            AWS_ACCOUNT_ID=${{ vars.CI_AWS_ACCOUNT_ID }}
            IMAGE_URI=${{ needs.build-sglang-image.outputs.ci-image }}
          else
            AWS_ACCOUNT_ID=${{ vars.PROD_AWS_ACCOUNT_ID }}
            IMAGE_URI=${{ vars.PROD_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_REGION }}.amazonaws.com/${{ env.PROD_SAGEMAKER_IMAGE }}
          fi

          echo "Image URI to test: ${IMAGE_URI}"
          echo "AWS_ACCOUNT_ID=${AWS_ACCOUNT_ID}" >> ${GITHUB_OUTPUT}
          echo "IMAGE_URI=${IMAGE_URI}" >> ${GITHUB_OUTPUT}

  sglang-telemetry-test:
    needs: [set-test-environment, build-sglang-image]
    if: success()
    runs-on:
      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
        buildspec-override:true
    concurrency:
      group: ${{ github.workflow }}-sglang-telemetry-test-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    steps:
      - name: Checkout DLC source
        uses: actions/checkout@v5

      - name: Container pull
        uses: ./.github/actions/ecr-authenticate
        with:
          aws-account-id: ${{ needs.set-test-environment.outputs.aws-account-id }}
          aws-region: ${{ vars.AWS_REGION }}
          image-uri: ${{ needs.set-test-environment.outputs.image-uri }}

      - name: Start container
        run: |
          CONTAINER_ID=$(docker run -d -it --rm --entrypoint /bin/bash \
            -e TEST_MODE=1 \
            ${{ needs.set-test-environment.outputs.image-uri }})
          echo "CONTAINER_ID=${CONTAINER_ID}" >> ${GITHUB_ENV}
          docker logs ${CONTAINER_ID}

      - name: Run tests
        run: |
          curl http://169.254.169.254/latest/meta-data/instance-id
          docker exec ${CONTAINER_ID} bash -c "export TEST_MODE=1 && python /usr/local/bin/deep_learning_container.py --framework sglang --framework-version "0.5.5" --container-type "general" && cat /tmp/test_request.txt"
          docker exec ${CONTAINER_ID} bash -c "export TEST_MODE=1 && bash /usr/local/bin/bash_telemetry.sh && sleep 30 && cat /tmp/test_request.txt | grep "${{ env.FRAMEWORK }}""


  sglang-local-benchmark-test:
    needs: [set-test-environment, build-sglang-image]
    if: success()
    runs-on:
      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
        fleet:x86-g6xl-runner
        buildspec-override:true
    concurrency:
      group: ${{ github.workflow }}-sglang-local-benchmark-test-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    steps:
      - name: Checkout DLC source
        uses: actions/checkout@v5

      - name: Container pull
        uses: ./.github/actions/ecr-authenticate
        with:
          aws-account-id: ${{ needs.set-test-environment.outputs.aws-account-id }}
          aws-region: ${{ vars.AWS_REGION }}
          image-uri: ${{ needs.set-test-environment.outputs.image-uri }}

      - name: Setup for SGLang datasets
        run: |
          mkdir -p ${TEST_ARTIFACTS_DIRECTORY}/dataset
          if [ ! -f ${TEST_ARTIFACTS_DIRECTORY}/dataset/ShareGPT_V3_unfiltered_cleaned_split.json ]; then
              echo "Downloading ShareGPT dataset..."
              wget -P ${TEST_ARTIFACTS_DIRECTORY}/dataset https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json
          else
              echo "ShareGPT dataset already exists. Skipping download."
          fi

      - name: Start container
        run: |
          CONTAINER_ID=$(docker run -d -it --rm --gpus=all \
            -v ${HOME}/.cache/huggingface:/root/.cache/huggingface \
            -v ${TEST_ARTIFACTS_DIRECTORY}/dataset:/dataset \
            -p 30000:30000 \
            -e SM_SGLANG_MODEL_PATH=Qwen/Qwen3-0.6B \
            -e SM_SGLANG_REASONING_PARSER=qwen3 \
            -e SM_SGLANG_HOST=127.0.0.1 \
            -e SM_SGLANG_PORT=30000 \
            -e HF_TOKEN=${{ secrets.HUGGING_FACE_HUB_TOKEN }} \
            ${{ needs.set-test-environment.outputs.image-uri }})
          echo "CONTAINER_ID=${CONTAINER_ID}" >> ${GITHUB_ENV}
          echo "Waiting for serving endpoint startup ..."
          sleep 120s
          docker logs ${CONTAINER_ID}

      - name: Run SGLang tests
        run: |
          docker exec ${CONTAINER_ID} python3 -m sglang.bench_serving \
          --backend sglang \
          --host 127.0.0.1 --port 30000 \
          --num-prompts 1000 \
          --model Qwen/Qwen3-0.6B \
          --dataset-name sharegpt \
          --dataset-path /dataset/ShareGPT_V3_unfiltered_cleaned_split.json

  sglang-frontend-test:
    needs: [build-sglang-image, set-test-environment]
    if: success()
    runs-on:
      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
        fleet:x86-g6exl-runner
        buildspec-override:true
    concurrency:
      group: ${{ github.workflow }}-sglang-frontend-test-${{ github.event.pull_request.number }}
      cancel-in-progress: true
    steps:
      - name: Checkout DLC source
        uses: actions/checkout@v5

      - name: Container pull
        uses: ./.github/actions/ecr-authenticate
        with:
          aws-account-id: ${{ needs.set-test-environment.outputs.aws-account-id }}
          aws-region: ${{ vars.AWS_REGION }}
          image-uri: ${{ needs.set-test-environment.outputs.image-uri }}

      - name: Checkout SGLang tests
        uses: actions/checkout@v5
        with:
          repository: sgl-project/sglang
          ref: v${{ env.SGLANG_VERSION }}
          path: sglang_source

      - name: Start container
        run: |
          CONTAINER_ID=$(docker run -d -it --rm --gpus=all --entrypoint /bin/bash \
            -v ${HOME}/.cache/huggingface:/root/.cache/huggingface \
            -v ./sglang_source:/workdir --workdir /workdir \
            -e HF_TOKEN=${{ secrets.HUGGING_FACE_HUB_TOKEN }} \
            ${{ needs.set-test-environment.outputs.image-uri }})
          echo "CONTAINER_ID=${CONTAINER_ID}" >> ${GITHUB_ENV}

      - name: Setup for SGLang tests
        run: |
          docker exec ${CONTAINER_ID} sh -c '
            set -eux

            bash scripts/ci/ci_install_dependency.sh
          '

      - name: Run SGLang tests
        run: |
          docker exec ${CONTAINER_ID} sh -c '
            set -eux
            nvidia-smi

            # Frontend Test
            cd /workdir/test/lang
            python3 run_suite.py --suite per-commit
          '

  sglang-sagemaker-endpoint-test:
    needs: [set-test-environment]
    if: |
      always() && !failure() && !cancelled() &&
      needs.set-test-environment.result == 'success'
    runs-on:
      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
        fleet:default-runner
        buildspec-override:true
    concurrency:
      group: ${{ github.workflow }}-sglang-sagemaker-endpoint-test-${{ github.event.pull_request.number }}
      cancel-in-progress: false
    steps:
      - name: Checkout DLC source
        uses: actions/checkout@v5

      - name: Install test dependencies
        run: |
          uv venv
          source .venv/bin/activate
          uv pip install -r test/requirements.txt
          uv pip install -r test/sglang/sagemaker/requirements.txt

      - name: Run sagemaker tests
        run: |
          source .venv/bin/activate
          cd test/
          python3 -m pytest -vs -rA --image-uri ${{ needs.set-test-environment.outputs.image-uri }} sglang/sagemaker