ci.yml

name: Process-agent CI

on:
  pull_request:
  push:
    branches:
      - master
  workflow_dispatch:

permissions: {}

concurrency:
  group: process-agent-ci-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

env:
  IMAGE: quay.io/stackstate/stackstate-k8s-process-agent

jobs:
  build-and-test:
    name: Prebuild, generated-code check, build, and unit tests (${{ matrix.arch }})
    runs-on: ${{ matrix.runner }}
    permissions:
      contents: read
    strategy:
      fail-fast: false
      matrix:
        include:
          - arch: amd64
            runner: ubuntu-24.04
            llvm-arch: x86_64
            prebuild-image: quay.io/stackstate/datadog_build_system-probe_x64:61b4ad67
            builder-image: quay.io/stackstate/datadog_build_deb_x64:61b4ad67
          - arch: arm64
            runner: ubuntu-24.04-arm
            llvm-arch: arm64
            prebuild-image: quay.io/stackstate/datadog_build_system-probe_arm64:61b4ad67
            builder-image: quay.io/stackstate/datadog_build_deb_arm64:61b4ad67
    steps:
      - name: Check out source commit
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          ref: ${{ github.event.pull_request.head.sha || github.sha }}
          fetch-depth: 0
          persist-credentials: false

      # A general-purpose BCI builder cannot generate DataDog's eBPF artifacts.
      # This repository-owned, architecture-specific builder is the existing
      # supported toolchain and is also used to build the upstream agent fork.
      - name: Generate DataDog eBPF and Go artifacts
        env:
          LLVM_ARCH: ${{ matrix.llvm-arch }}
          PREBUILD_IMAGE: ${{ matrix.prebuild-image }}
        run: |
          set -euo pipefail
          docker run --rm \
            --volume "${GITHUB_WORKSPACE}:/workspace" \
            --workdir /workspace \
            --env LLVM_ARCH="${LLVM_ARCH}" \
            --env OUTPUT_USER_ID="$(id -u)" \
            --env OUTPUT_GROUP_ID="$(id -g)" \
            "${PREBUILD_IMAGE}" \
            bash -c '
              set -euo pipefail
              ./prebuild-datadog-agent.sh --generate-no-docker
            '

      # The DataDog builder carries the matching compiler, Ruby, conda, LLVM,
      # and native libraries needed by the fork. There is no equivalent BCI
      # image that can compile this legacy eBPF dependency graph.
      - name: Verify generated code, build, and test
        env:
          BUILDER_IMAGE: ${{ matrix.builder-image }}
          SOURCE_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
        run: |
          set -euo pipefail
          short_sha="$(printf '%s' "${SOURCE_SHA}" | cut -c1-8)"

          docker run --rm \
            --volume "${GITHUB_WORKSPACE}:/workspace" \
            --workdir /workspace \
            --env CI=true \
            --env GO111MODULE=on \
            --env GOPATH=/workspace/.go \
            --env PROCESS_AGENT_VERSION="${short_sha}" \
            "${BUILDER_IMAGE}" \
            bash -c '
              source /root/.bashrc
              conda activate ddpy3
              # The RVM directory hook reads an unset rvm_saved_env variable.
              # Keep fail-fast/pipefail without nounset in this legacy builder.
              set -eo pipefail
              export PATH="${GOPATH}/bin:${PATH}"

              go mod verify
              gogo_version="$(go list -m -f "{{.Version}}" github.com/gogo/protobuf)"
              go install "github.com/gogo/protobuf/protoc-gen-gogofaster@${gogo_version}"
              rake protobuf
              git diff --exit-code -- model
              rake ci
              ./prebuild-datadog-agent.sh --install-ebpf
              git diff --exit-code
            '

      - name: Upload process-agent image inputs
        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
        with:
          name: process-agent-${{ matrix.arch }}
          path: |
            process-agent
            ebpf-object-files
          if-no-files-found: error
          retention-days: 1

  image-smoke-and-scan:
    name: BCI image smoke test, Trivy secrets/CVEs, and Grype (${{ matrix.arch }})
    needs: build-and-test
    runs-on: ${{ matrix.runner }}
    permissions:
      contents: read
    strategy:
      fail-fast: false
      matrix:
        include:
          - arch: amd64
            runner: ubuntu-24.04
            long-arch: x86_64
            llvm-arch: x86_64
          - arch: arm64
            runner: ubuntu-24.04-arm
            long-arch: aarch64
            llvm-arch: arm64
    steps:
      - name: Check out source commit
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          ref: ${{ github.event.pull_request.head.sha || github.sha }}
          persist-credentials: false

      - name: Download process-agent image inputs
        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
        with:
          name: process-agent-${{ matrix.arch }}
          path: .

      - name: Set up Docker Buildx
        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0

      - name: Resolve image metadata
        id: image
        env:
          SOURCE_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
        run: |
          set -euo pipefail
          short_sha="$(printf '%s' "${SOURCE_SHA}" | cut -c1-8)"
          echo "tag=${short_sha}" >> "${GITHUB_OUTPUT}"

      - name: Build local BCI runtime image
        uses: docker/build-push-action@bcafcacb16a39f128d818304e6c9c0c18556b85f # v7.1.0
        with:
          context: .
          file: BCI.dockerfile
          platforms: linux/${{ matrix.arch }}
          load: true
          push: false
          provenance: false
          sbom: false
          build-args: |
            EBPF_SUBFOLDER=${{ matrix.llvm-arch }}
            LONG_ARCH=${{ matrix.long-arch }}
            SHORT_ARCH=${{ matrix.arch }}
          tags: ${{ env.IMAGE }}:${{ steps.image.outputs.tag }}-${{ matrix.arch }}

      - name: Smoke test packaged process-agent binary
        env:
          ARCH: ${{ matrix.arch }}
          TAG: ${{ steps.image.outputs.tag }}
        run: |
          set -euo pipefail
          image="${IMAGE}:${TAG}-${ARCH}"
          output="$(docker run --rm \
            --entrypoint /opt/stackstate-agent/bin/agent/process-agent \
            "${image}" -version)"
          printf '%s\n' "${output}"
          grep -F "Version: ${TAG}" <<< "${output}"

          healthcheck="$(docker image inspect --format '{{json .Config.Healthcheck.Test}}' "${image}")"
          grep -F '/probe.sh' <<< "${healthcheck}"

      - name: Scan image with VEX-aware Trivy and Grype
        uses: StackVista/image-pipeline/.github/actions/scan-image@6284a6fc006a7cc46a7f00d02c50d5f21b117b63
        with:
          image: ${{ env.IMAGE }}:${{ steps.image.outputs.tag }}-${{ matrix.arch }}
          mode: gate
          severity: UNKNOWN,LOW,MEDIUM,HIGH,CRITICAL
          with-grype: true
          exceptions-path: exceptions
          upload-sarif: false
          sarif-category: process-agent-${{ matrix.arch }}

  publish-image:
    name: Publish and sign commit image (${{ matrix.arch }})
    needs: image-smoke-and-scan
    if: >-
      (github.event_name == 'pull_request' &&
      github.event.pull_request.head.repo.full_name == github.repository) ||
      ((github.event_name == 'push' || github.event_name == 'workflow_dispatch') &&
      github.ref == 'refs/heads/master')
    runs-on: ${{ matrix.runner }}
    permissions:
      contents: read
      id-token: write
    strategy:
      fail-fast: false
      matrix:
        include:
          - arch: amd64
            runner: ubuntu-24.04
            long-arch: x86_64
            llvm-arch: x86_64
          - arch: arm64
            runner: ubuntu-24.04-arm
            long-arch: aarch64
            llvm-arch: arm64
    steps:
      - name: Check out source commit
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          ref: ${{ github.event.pull_request.head.sha || github.sha }}
          persist-credentials: false

      - name: Download process-agent image inputs
        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
        with:
          name: process-agent-${{ matrix.arch }}
          path: .

      - name: Set up Docker Buildx
        uses: docker/setup-buildx-action@4d04d5d9486b7bd6fa91e7baf45bbb4f8b9deedd # v4.0.0

      - name: Resolve image tag
        id: image
        env:
          SOURCE_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
        run: |
          set -euo pipefail
          short_sha="$(printf '%s' "${SOURCE_SHA}" | cut -c1-8)"
          echo "tag=${short_sha}" >> "${GITHUB_OUTPUT}"

      - name: Resolve canonical OCI labels
        id: oci
        uses: StackVista/image-pipeline/.github/actions/apply-oci-labels@6284a6fc006a7cc46a7f00d02c50d5f21b117b63
        with:
          image-name: stackstate-k8s-process-agent
          tag: ${{ steps.image.outputs.tag }}
          title: SUSE Observability Process Agent
          description: Process agent collecting per-process and per-container telemetry for SUSE Observability.
          component: stackstate-k8s-process-agent
          dockerfile: BCI.dockerfile
          base-name: registry.suse.com/bci/bci-micro:15.7

      - name: Build, publish, and sign architecture image
        uses: StackVista/image-pipeline/.github/actions/push-single-arch@6284a6fc006a7cc46a7f00d02c50d5f21b117b63
        with:
          image: ${{ env.IMAGE }}
          tag: ${{ steps.image.outputs.tag }}
          arch: ${{ matrix.arch }}
          dockerfile: BCI.dockerfile
          labels: |
            ${{ steps.oci.outputs.labels }}
            org.opencontainers.image.revision=${{ github.event.pull_request.head.sha || github.sha }}
          build-args: |
            EBPF_SUBFOLDER=${{ matrix.llvm-arch }}
            LONG_ARCH=${{ matrix.long-arch }}
            SHORT_ARCH=${{ matrix.arch }}
          target-registry: quay.io
          target-registry-user: ${{ vars.QUAY_USER }}
          target-registry-password: ${{ secrets.QUAY_PASSWORD }}

  merge-multiarch-manifest:
    name: Publish and sign multi-architecture commit image
    needs: publish-image
    if: >-
      (github.event_name == 'pull_request' &&
      github.event.pull_request.head.repo.full_name == github.repository) ||
      ((github.event_name == 'push' || github.event_name == 'workflow_dispatch') &&
      github.ref == 'refs/heads/master')
    runs-on: ubuntu-24.04
    permissions:
      contents: read
      id-token: write
    steps:
      - name: Resolve image tag
        id: image
        env:
          SOURCE_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
        run: |
          set -euo pipefail
          short_sha="$(printf '%s' "${SOURCE_SHA}" | cut -c1-8)"
          echo "tag=${short_sha}" >> "${GITHUB_OUTPUT}"

      - name: Merge and sign multi-architecture manifest
        uses: StackVista/image-pipeline/.github/actions/merge-multiarch@6284a6fc006a7cc46a7f00d02c50d5f21b117b63
        with:
          image: ${{ env.IMAGE }}
          tag: ${{ steps.image.outputs.tag }}
          target-registry: quay.io
          target-registry-user: ${{ vars.QUAY_USER }}
          target-registry-password: ${{ secrets.QUAY_PASSWORD }}

  ci-success:
    name: Process-agent CI
    needs:
      - build-and-test
      - image-smoke-and-scan
      - publish-image
      - merge-multiarch-manifest
    if: always()
    runs-on: ubuntu-24.04
    permissions: {}
    steps:
      - name: Verify all required jobs succeeded
        env:
          NEEDS: ${{ toJSON(needs) }}
        run: |
          set -euo pipefail
          failed="$(jq -r '
            to_entries[] |
            select(.value.result != "success" and .value.result != "skipped") |
            .key
          ' <<< "${NEEDS}")"
          if [ -n "${failed}" ]; then
            echo "Required process-agent jobs failed:"
            printf '%s\n' "${failed}"
            exit 1
          fi
          echo "All required process-agent jobs passed."

  # Terminal job, master only. ci-success aggregates every other job, so hanging
  # the notification off it means one funnel for all failures rather than a
  # notify job per pipeline job. Replaces `notify-on-master-fail` from the
  # retired .gitlab-ci.yml (STAC-25519).
  cerberus-notify:
    name: Report failure to Slack (Cerberus)
    needs: ci-success
    if: >-
      failure() &&
      github.ref == 'refs/heads/master' &&
      github.event_name == 'push'
    uses: ./.github/workflows/cerberus-notify.yml
    with:
      suite: build
    secrets:
      CERBERUS_LAMBDA_URL: ${{ secrets.CERBERUS_LAMBDA_URL }}
      CERBERUS_API_TOKEN: ${{ secrets.CERBERUS_API_TOKEN }}