ci-coordinator.yaml

name: CI - Coordinator Test

# Runs coordinator unit tests and e2e tests. Path filter excludes paths not
# exercised by the coordinator. The e2e-images and e2e-tests matrix jobs are
# each followed by a *-status job that aggregates the matrix result, since
# GitHub required status checks cannot reference individual matrix legs.
on:
  push:
    branches:
      - main
      - 'release-*'
  pull_request:
    branches:
      - main

permissions:
  contents: read
  actions: read

concurrency:
  group: ${{ github.workflow }}-${{ github.event_name }}-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

jobs:
  check-changes:
    runs-on: ubuntu-latest
    outputs:
      src: ${{ steps.filter.outputs.src }}
    steps:
      - name: Checkout source
        uses: actions/checkout@v7
      - uses: dorny/paths-filter@v4
        id: filter
        with:
          filters: |
            src:
              - '**/*.go'
              - Dockerfile.*
              - Makefile*
              - go.mod
              - go.sum
              - scripts/**
              - hack/**
              - test/**
              - config/coordinator/**
              - deploy/**
              - .github/actions/docker-build-and-push/action.yml
              - .github/actions/e2e-runner-setup/action.yml
              - .github/workflows/ci-coordinator.yaml
              - '!Dockerfile.sidecar'
              - '!cmd/pd-sidecar/**'
              - '!pkg/sidecar/**'
              - '!test/sidecar/**'
              - '!pkg/generator/**'
              - '!test/e2e/**'
              - '!test/integration/**'
              - '!test/profiling/**'
              - '!Makefile'

  unit-tests:
    needs: check-changes
    if: ${{ needs.check-changes.outputs.src == 'true' }}
    runs-on: ubuntu-latest
    timeout-minutes: 30
    steps:
      - name: Checkout source
        uses: actions/checkout@v7

      - name: Create Go cache dirs
        id: go-cache
        run: |
          mkdir -p "$HOME/.cache/llm-d-gomodcache" "$HOME/.cache/llm-d-gobuildcache"
          echo "mod=$HOME/.cache/llm-d-gomodcache" >> "$GITHUB_OUTPUT"
          echo "build=$HOME/.cache/llm-d-gobuildcache" >> "$GITHUB_OUTPUT"

      - name: Cache Go modules and build cache
        uses: actions/cache@v6
        with:
          path: |
            ${{ steps.go-cache.outputs.mod }}
            ${{ steps.go-cache.outputs.build }}
          key: go-cache-${{ runner.os }}-${{ hashFiles('go.sum') }}
          restore-keys: |
            go-cache-${{ runner.os }}-

      - name: Run coordinator unit tests
        shell: bash
        env:
          GO_MOD_CACHE_VOL: ${{ steps.go-cache.outputs.mod }}
          GO_BUILD_CACHE_VOL: ${{ steps.go-cache.outputs.build }}
        run: make -f Makefile.coord.mk test-unit

  e2e-images:
    needs: check-changes
    if: ${{ needs.check-changes.outputs.src == 'true' }}
    runs-on: ubuntu-latest
    timeout-minutes: 20
    strategy:
      fail-fast: false
      matrix:
        image:
          - name: builder
            source: build
            docker-file: Dockerfile.builder
            image-name: llm-d-builder
            archive: llm-d-builder
          - name: coordinator
            source: build
            docker-file: Dockerfile.coordinator
            image-name: llm-d-coordinator
            archive: llm-d-coordinator
          - name: epp
            source: build
            docker-file: Dockerfile.epp
            image-name: llm-d-router-endpoint-picker
            archive: llm-d-router-endpoint-picker
          - name: simulator
            source: pull
            image: ghcr.io/llm-d/llm-d-inference-sim:v0.10.2
            archive: llm-d-inference-sim
    name: e2e-images (${{ matrix.image.name }})
    steps:
      - name: Checkout source
        if: ${{ matrix.image.source == 'build' }}
        uses: actions/checkout@v7

      - name: Build ${{ matrix.image.name }} image
        if: ${{ matrix.image.source == 'build' }}
        uses: ./.github/actions/docker-build-and-push
        with:
          docker-file: ${{ matrix.image.docker-file }}
          image-name: ${{ matrix.image.image-name }}
          tag: dev
          registry: ghcr.io/llm-d
          push: 'false'
          buildx-outputs: type=docker,dest=${{ runner.temp }}/${{ matrix.image.archive }}.tar
          commit-sha: ${{ github.sha }}
          build-ref: ${{ github.head_ref || github.ref_name }}

      - name: Pull ${{ matrix.image.name }} image
        if: ${{ matrix.image.source == 'pull' }}
        run: |
          docker pull --platform linux/amd64 "${{ matrix.image.image }}"
          docker save "${{ matrix.image.image }}" -o "${{ runner.temp }}/${{ matrix.image.archive }}.tar"

      - name: Upload ${{ matrix.image.name }} image
        uses: actions/upload-artifact@v7
        with:
          name: e2e-coord-image-${{ matrix.image.name }}
          path: ${{ runner.temp }}/${{ matrix.image.archive }}.tar
          retention-days: 1
          compression-level: 0

  e2e-images-status:
    needs: e2e-images
    if: always()
    runs-on: ubuntu-latest
    steps:
      - name: Check e2e-images result
        run: |
          result="${{ needs.e2e-images.result }}"
          if [[ "$result" != "success" && "$result" != "skipped" ]]; then
            echo "::error::e2e-images failed for at least one matrix image"
            exit 1
          fi

  e2e-tests:
    needs:
      - check-changes
      - e2e-images
    if: ${{ needs.check-changes.outputs.src == 'true' }}
    runs-on: ubuntu-latest
    timeout-minutes: 60
    strategy:
      fail-fast: false
      matrix:
        include:
          - topology: single-EPP
            target: test-e2e-coordinator-run
          - topology: 3-EPP
            target: test-e2e-coordinator-3epp-run
    name: e2e-tests (${{ matrix.topology }})
    steps:
      - name: Checkout source
        uses: actions/checkout@v7

      - name: Set up e2e runner
        id: e2e-setup
        uses: ./.github/actions/e2e-runner-setup

      - name: Download builder image
        uses: actions/download-artifact@v8
        with:
          name: e2e-coord-image-builder
          path: ${{ runner.temp }}/e2e-images

      - name: Download coordinator image
        uses: actions/download-artifact@v8
        with:
          name: e2e-coord-image-coordinator
          path: ${{ runner.temp }}/e2e-images

      - name: Download EPP image
        uses: actions/download-artifact@v8
        with:
          name: e2e-coord-image-epp
          path: ${{ runner.temp }}/e2e-images

      - name: Download simulator image
        uses: actions/download-artifact@v8
        with:
          name: e2e-coord-image-simulator
          path: ${{ runner.temp }}/e2e-images

      - name: Load e2e images
        shell: bash
        run: |
          for image in "${RUNNER_TEMP}"/e2e-images/*.tar; do
            docker load --input "$image"
          done

      - name: Run coordinator e2e tests
        shell: bash
        env:
          GO_MOD_CACHE_VOL: ${{ steps.e2e-setup.outputs.go-mod-cache }}
          GO_BUILD_CACHE_VOL: ${{ steps.e2e-setup.outputs.go-build-cache }}
        run: make -f Makefile.coord.mk ${{ matrix.target }}

  e2e-tests-status:
    needs: e2e-tests
    if: always()
    runs-on: ubuntu-latest
    steps:
      - name: Check e2e-tests result
        run: |
          result="${{ needs.e2e-tests.result }}"
          if [[ "$result" != "success" && "$result" != "skipped" ]]; then
            echo "::error::e2e-tests failed for at least one topology"
            exit 1
          fi