run_spicebench.yml

name: Run
run-name: Run - ${{ github.event.inputs.system_under_test || 'spice_cloud' }} - ${{ github.event.inputs.etl_type }} - SF${{ github.event.inputs.scale_factor }}

on:
  schedule:
    - cron: '0 6 * * *'  # Daily SF1.0 SCP run
  workflow_dispatch:
    inputs:
      scenario:
        description: 'Scenario/query set to run'
        required: true
        default: 'tpch'
        type: choice
        options:
          - tpch
      system_under_test:
        description: 'System under test'
        required: true
        default: spice_cloud
        type: choice
        options:
          - spice_cloud
          - databricks-sql
          - databricks-lakebase
      etl_type:
        description: 'ETL type'
        required: true
        default: 'changes'
        type: choice
        options:
          - events
          - changes
      scale_factor:
        description: 'Scale Factor'
        required: true
        default: '1'
        type: choice
        options:
          - '0.1'
          - '1'
          - '10'

env:
  SCENARIO: ${{ github.event.inputs.scenario || 'tpch' }}
  SYSTEM_UNDER_TEST: ${{ github.event.inputs.system_under_test || 'spice_cloud' }}
  ETL_TYPE: ${{ github.event.inputs.etl_type || 'changes' }}
  SCALE_FACTOR: ${{ github.event.inputs.scale_factor || '1' }}

jobs:
  run-spicebench:
    name: Run spicebench
    runs-on: spiceai-dev-runners
    timeout-minutes: 600
    steps:
      - uses: actions/checkout@v6

      - uses: ./.github/actions/setup-cc

      - uses: ./.github/actions/management-login
        if: ${{ env.SYSTEM_UNDER_TEST == 'spice_cloud' }}
        with:
          token-url: https://spice.ai/api/oauth/token
          client-id: ${{ secrets.SPICE_MANAGEMENT_CLIENT_ID_PROD }}
          client-secret: ${{ secrets.SPICE_MANAGEMENT_CLIENT_SECRET_PROD }}

      - name: Log in to GHCR
        if: ${{ env.SYSTEM_UNDER_TEST == 'spice_cloud' }}
        uses: docker/login-action@v3
        with:
          registry: ghcr.io
          username: ${{ github.actor }}
          password: ${{ secrets.GITHUB_TOKEN }}

      - name: Pull spidapter image
        if: ${{ env.SYSTEM_UNDER_TEST == 'spice_cloud' }}
        run: docker pull ghcr.io/spiceai/spidapter:latest

      - uses: ./.github/actions/build-spicebench

      - name: Restore databricks adapter cache
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') }}
        id: cache-databricks-adapter
        uses: actions/cache/restore@v4
        with:
          path: ~/.spice/bin/databricks-system-adapter
          key: databricks-system-adapter-${{ runner.os }}-${{ hashFiles('system-adapters/databricks/Cargo.toml', 'system-adapters/databricks/Cargo.lock', 'system-adapters/databricks/src/**/*.rs', 'crates/system-adapter-protocol/Cargo.toml', 'crates/system-adapter-protocol/src/**/*.rs') }}
          restore-keys: |
            databricks-system-adapter-${{ runner.os }}-

      - name: Build databricks adapter
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') && steps.cache-databricks-adapter.outputs.cache-hit != 'true' }}
        id: build-databricks-adapter
        run: |
          mkdir -p ~/.spice/bin
          cargo build --manifest-path system-adapters/databricks/Cargo.toml
          install -m 755 system-adapters/databricks/target/debug/databricks-system-adapter ~/.spice/bin/databricks-system-adapter

      - name: Save databricks adapter cache
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') && steps.build-databricks-adapter.outcome == 'success' }}
        uses: actions/cache/save@v4
        with:
          path: ~/.spice/bin/databricks-system-adapter
          key: databricks-system-adapter-${{ runner.os }}-${{ hashFiles('system-adapters/databricks/Cargo.toml', 'system-adapters/databricks/Cargo.lock', 'system-adapters/databricks/src/**/*.rs', 'crates/system-adapter-protocol/Cargo.toml', 'crates/system-adapter-protocol/src/**/*.rs') }}

      - name: Setup Go
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') }}
        uses: actions/setup-go@v5
        with:
          go-version: '1.23'

      - name: Checkout adbc-databricks Go driver
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') }}
        uses: actions/checkout@v6
        with:
          repository: spiceai/adbc-databricks
          ref: spicebench
          path: adbc-databricks

      - name: Build databricks Go ADBC driver
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') }}
        run: |
          cd adbc-databricks/go
          go build -tags driverlib -buildmode=c-shared \
            -o build/libadbc_driver_databricks.so \
            ./pkg/
          sudo install -m 755 build/libadbc_driver_databricks.so /usr/local/lib/libdatabricks.so
          sudo ldconfig

      - name: Install ADBC Postgres driver
        if: ${{ startsWith(env.SYSTEM_UNDER_TEST, 'databricks-') }}
        uses: columnar-tech/setup-dbc@v1
        with:
          drivers: postgresql

      - name: Install ADBC FlightSQL driver
        if: ${{ env.SYSTEM_UNDER_TEST == 'spice_cloud' }}
        uses: columnar-tech/setup-dbc@v1
        with:
          drivers: flightsql

      - name: Run spicebench
        env:
          SPICEAI_API_KEY: ${{ env.SPICEAI_API_KEY }}
          SPICE_CLOUD_API_URL: https://api.spice.ai
          DATABRICKS_ENDPOINT: ${{ secrets.DATABRICKS_ENDPOINT }}
          DATABRICKS_TOKEN: ${{ secrets.DATABRICKS_TOKEN }}
          DATABRICKS_HTTP_PATH: ${{ secrets.DATABRICKS_HTTP_PATH }}
          DATABRICKS_SQL_WAREHOUSE_ID: ${{ secrets.DATABRICKS_SQL_WAREHOUSE_ID }}
          DATABRICKS_CATALOG: ${{ vars.DATABRICKS_CATALOG }}
          DATABRICKS_SCHEMA: ${{ vars.DATABRICKS_SCHEMA }}
          DATABRICKS_STAGING_VOLUME_PATH: ${{ vars.DATABRICKS_STAGING_VOLUME_PATH }}
          LAKEBASE_PG_HOST: ${{ vars.LAKEBASE_PG_HOST }}
          LAKEBASE_PG_USER: ${{ vars.LAKEBASE_PG_USER }}
          LAKEBASE_PG_SCHEMA: ${{ vars.LAKEBASE_PG_SCHEMA }}
          LAKEBASE_PROJECT: ${{ vars.LAKEBASE_PROJECT }}
          LAKEBASE_BRANCH: ${{ vars.LAKEBASE_BRANCH }}
          SYSTEM_ADAPTER: ${{ env.SYSTEM_UNDER_TEST }}
          NUM_QUERY_CLIENTS: '2'
          ETL_BUCKET: spicebench
          ETL_PREFIX: ${{ env.ETL_TYPE == 'changes' && 'data-gen-mutable' || 'data-gen' }}
          ETL_REGION: us-east-1
          ETL_SINK: adbc
          SCHEDULER_STATE_LOCATION: s3://spiceai-testing-cluster-state/spicebench-scheduler-state-${{ github.run_id }}/
          VALIDATE_CHECKPOINT_RESULTS: 'true'
          SCRAPE_SUT_METRICS: 'true'
          SPICEAI_BENCHMARK_METRICS_KEY: ${{ secrets.SPICEAI_BENCHMARK_METRICS_KEY }}
          MINIO_ENDPOINT: ${{ secrets.MINIO_ENDPOINT }}
          AWS_ACCESS_KEY_ID: ${{ secrets.MINIO_ACCESS_KEY_ID }}
          AWS_SECRET_ACCESS_KEY: ${{ secrets.MINIO_SECRET_ACCESS_KEY }}
          S3_AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
          S3_AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          SPIDAPTER_APP_MEMORY_LIMIT: '62Gi'
          SPIDAPTER_EXECUTOR_MEMORY_LIMIT: '62Gi'
          SPIDAPTER_EPHEMERAL_STORAGE_LIMIT_GB: '256Gi'
          SPIDAPTER_ORGANIZATION_TAG: spicehq
          SPIDAPTER_ICEBERG_REGION: us-west-1
          SPIDAPTER_ICEBERG_CATALOG_FROM: iceberg:https://glue.us-west-1.amazonaws.com/iceberg/v1/catalogs/211125479522/namespaces
        run: |
          set -euo pipefail
          export RUST_LOG='info,etl::sink::adbc=debug'

          EXECUTOR_INSTANCE_TYPE="github-hosted-ubuntu-latest"
          ETL_ENDPOINT="${MINIO_ENDPOINT}"

          ETL_ARGS="--etl-bucket ${ETL_BUCKET} --scale-factor ${SCALE_FACTOR} --etl-prefix ${ETL_PREFIX} --etl-region ${ETL_REGION}"
          [ -n "${ETL_ENDPOINT:-}" ] && ETL_ARGS="${ETL_ARGS} --etl-endpoint ${ETL_ENDPOINT}"

          VALIDATION_ARGS=""
          [ "${VALIDATE_CHECKPOINT_RESULTS}" = "true" ] && VALIDATION_ARGS="--validate-results"

          SUT_METRICS_ARGS=""
          [ "${SCRAPE_SUT_METRICS}" = "true" ] && SUT_METRICS_ARGS="--scrape-sut-metrics"

          if [ "${SYSTEM_UNDER_TEST}" = "spice_cloud" ]; then
            export SPICEBENCH_ADBC_UPDATE_STRATEGY=bulk_ingest_upsert
            export SPICEBENCH_ADBC_FLUSH_STREAM_BEFORE_UPSERT=true
            export SPICEBENCH_ADBC_DELETE_BATCH_SIZE=50000

            ADAPTER_DOCKER_OPTS="run -i"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_SCENARIO=cayenne"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPICEAI_API_KEY -e SPICE_CLOUD_API_URL"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e AWS_ACCESS_KEY_ID=${S3_AWS_ACCESS_KEY_ID} -e AWS_SECRET_ACCESS_KEY=${S3_AWS_SECRET_ACCESS_KEY}"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SCHEDULER_STATE_LOCATION"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_ORGANIZATION_TAG"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_APP_MEMORY_LIMIT"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_EXECUTOR_MEMORY_LIMIT"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_EPHEMERAL_STORAGE_LIMIT_GB"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_ICEBERG_REGION"
            ADAPTER_DOCKER_OPTS="${ADAPTER_DOCKER_OPTS} -e SPIDAPTER_ICEBERG_CATALOG_FROM"

            ADAPTER_CMD="docker"
            ADAPTER_ARGS="${ADAPTER_DOCKER_OPTS} ghcr.io/spiceai/spidapter:latest stdio --verbose"
            ADAPTER_ENVS="--system-adapter-env SCHEDULER_STATE_LOCATION=${SCHEDULER_STATE_LOCATION}"

          else
            # databricks-sql / databricks-lakebase
            export SPICEBENCH_ADBC_DELETE_BATCH_SIZE=50000
            export SPICEBENCH_ADBC_UPDATE_STRATEGY=staging_table
            export SPICEBENCH_TARGET_BATCH_ROWS=500000
            # ADBC writes are heavyweight (staging-parquet upload + MERGE), so
            # splitting a segment into many small writes hurts: disable chunking
            # (one write per segment) and let each table do 1 in-flight write with
            # no global ceiling (tables ingest concurrently). Mirrors trunk.
            export SPICEBENCH_SINK_CHUNK_ROWS=0
            export SPICEBENCH_SINK_PARALLELISM_PER_TABLE=1
            export SPICEBENCH_ADBC_MAX_INGEST_BATCH_BYTES=1268435456
            export SPICEBENCH_ADBC_REUSE_BULK_INGEST_STREAMS=false
            export SPICEBENCH_ADBC_ANALYZE_STAGING_BEFORE_MERGE=true

            ADAPTER_CMD="${HOME}/.spice/bin/databricks-system-adapter"
            ADAPTER_ARGS="stdio"
            ADAPTER_ENVS="--system-adapter-env DATABRICKS_ENDPOINT=${DATABRICKS_ENDPOINT}"
            ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_TOKEN=${DATABRICKS_TOKEN}"
            ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_HTTP_PATH=${DATABRICKS_HTTP_PATH}"
            ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_SQL_WAREHOUSE_ID=${DATABRICKS_SQL_WAREHOUSE_ID}"
            ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_TABLE_FORMAT=parquet"
            [ -n "${DATABRICKS_CATALOG:-}" ]             && ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_CATALOG=${DATABRICKS_CATALOG}"
            [ -n "${DATABRICKS_SCHEMA:-}" ]              && ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_SCHEMA=${DATABRICKS_SCHEMA}"
            [ -n "${DATABRICKS_STAGING_VOLUME_PATH:-}" ] && ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_STAGING_VOLUME_PATH=${DATABRICKS_STAGING_VOLUME_PATH}"

            if [ "${SYSTEM_UNDER_TEST}" = "databricks-lakebase" ]; then
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env DATABRICKS_COMPUTE_MODE=lakebase"
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env LAKEBASE_PG_HOST=${LAKEBASE_PG_HOST}"
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env LAKEBASE_PG_USER=${LAKEBASE_PG_USER}"
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env LAKEBASE_PG_SCHEMA=${LAKEBASE_PG_SCHEMA}"
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env LAKEBASE_PROJECT=${LAKEBASE_PROJECT}"
              ADAPTER_ENVS="${ADAPTER_ENVS} --system-adapter-env LAKEBASE_BRANCH=${LAKEBASE_BRANCH}"
            fi
          fi

          set -x
          ~/.spice/bin/spicebench run \
            --concurrency "${NUM_QUERY_CLIENTS}" \
            --scenario "${SCENARIO}" \
            --executor-instance-type "${EXECUTOR_INSTANCE_TYPE}" \
            ${ETL_ARGS} \
            --etl-sink "${ETL_SINK}" \
            --table-format parquet \
            ${VALIDATION_ARGS} \
            ${SUT_METRICS_ARGS} \
            --system-adapter-stdio-cmd "${ADAPTER_CMD}" \
            --system-adapter-stdio-args "${ADAPTER_ARGS}" \
            ${ADAPTER_ENVS}