Verwendung von Google Cloud-Funktionen für die dreistufige Datenverarbeitung mit Google Composer und automatisierten Bereitstellungen über GitHub-Aktionen

Inhaltsverzeichnis

1. Über
2. Verwendete Werkzeuge
3. Lösungsarchitekturdiagramm
4. Bereitstellungsprozess
   • Voraussetzungen
   • Geheimnisse für GitHub-Aktionen
   • So erstellen Sie ein neues Geheimnis:
   • Einrichten des DevOps-Dienstkontos
5. GitHub Actions Pipeline: Schritte
   • Enable-Services
   • deploy-buckets
   • cloud-funktion bereitstellen
   • deploy-composer-service-account
   • deploy-bigquery-dataset-bigquery-tables
   • deploy-composer-environment
   • deploy-composer-http-connection
   • deploy-dags
6. GitHub Actions Workflow-Erklärung
7. Nach der Bereitstellung erstellte Ressourcen
8. Fazit
9. Referenzen

Um

In diesem Beitrag wird die Verwendung von Google Cloud Functions für die Datenverarbeitung in einer dreistufigen Architektur untersucht. Die Lösung ist mit Google Composer orchestriert und bietet automatisierte Bereitstellungen mithilfe von GitHub Actions. Wir gehen durch die verwendeten Tools, den Bereitstellungsprozess und die Pipeline-Schritte und bieten einen klaren Leitfaden für den Aufbau einer durchgängigen cloudbasierten Datenpipeline.

Verwendete Werkzeuge

1. Google Cloud Platform (GCP): Die primäre Cloud-Umgebung.
2. Cloud-Speicher: Zum Speichern eingegebener und verarbeiteter Daten über verschiedene Ebenen hinweg (Bronze, Silber, Gold).
3. Cloud-Funktionen: Serverlose Funktionen, die für die Datenverarbeitung in jeder Ebene verantwortlich sind.
4. Google Composer: Ein auf Apache Airflow basierendes Orchestrierungstool, das zum Planen und Verwalten von Arbeitsabläufen verwendet wird.
5. GitHub-Aktionen:Automatisierungstool für die Bereitstellung und Verwaltung der Pipeline-Infrastruktur.

Lösungsarchitekturdiagramm

Bereitstellungsprozess

Voraussetzungen

Bevor Sie das Projekt einrichten, stellen Sie sicher, dass Sie Folgendes haben:

1. GCP-Konto: Ein Google Cloud-Konto mit aktivierter Abrechnung.
2. Dienstkonto für DevOps: Ein Dienstkonto mit den erforderlichen Berechtigungen zum Bereitstellen von Ressourcen in GCP.
3. Geheimnisse in GitHub: Speichern Sie die Anmeldeinformationen des GCP-Dienstkontos, die Projekt-ID und den Bucket-Namen als Geheimnisse in GitHub für sicheren Zugriff.

Geheimnisse für GitHub-Aktionen

Um sicher auf Ihr GCP-Projekt und Ihre GCP-Ressourcen zuzugreifen, legen Sie die folgenden Geheimnisse in GitHub-Aktionen fest:

• BUCKET_DATALAKE: Ihr Cloud Storage-Bucket für den Data Lake.
• GCP_DEVOPS_SA_KEY: Der Dienstkontoschlüssel im JSON-Format.
• PROJECT_ID: Ihre GCP-Projekt-ID.
• REGION_PROJECT_ID: Die Region, in der Ihr GCP-Projekt bereitgestellt wird.

So erstellen Sie ein neues Geheimnis:

1. In project repository, menu **Settings** 
2. **Security**, 
3. **Secrets and variables**,click in access **Action**
4. **New repository secret**, type a **name** and **value** for secret.

Weitere Informationen finden Sie unter:
https://docs.github.com/pt/actions/security-for-github-actions/security-guides/using-secrets-in-github-actions

Einrichten des DevOps-Dienstkontos

Erstellen Sie ein Dienstkonto in GCP mit Berechtigungen für Cloud Functions, Composer, BigQuery und Cloud Storage. Gewähren Sie die erforderlichen Rollen, z. B.:

• Cloud Functions-Administrator
• Komponistenbenutzer
• BigQuery-Dateneditor
• Speicherobjektverwaltung

GitHub Actions Pipeline: Schritte

Die Pipeline automatisiert den gesamten Bereitstellungsprozess und stellt sicher, dass alle Komponenten korrekt eingerichtet sind. Hier ist eine Aufschlüsselung der wichtigsten Jobs aus der GitHub-Aktionsdatei, von denen jeder für einen anderen Aspekt der Bereitstellung verantwortlich ist.

Enable-Dienste

enable-services:
    runs-on: ubuntu-22.04
    steps:
    - uses: actions/checkout@v2

    # Step to Authenticate with GCP
    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Configure  Cloud SDK
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker

    - name: Set up python 3.8
      uses: actions/setup-python@v2
      with:
        python-version: 3.8.16

    # Step to Create GCP Bucket 
    - name: Enable gcp service api's
      run: |-
        gcloud services enable ${{ env.GCP_SERVICE_API_0 }}
        gcloud services enable ${{ env.GCP_SERVICE_API_1 }}
        gcloud services enable ${{ env.GCP_SERVICE_API_2 }}
        gcloud services enable ${{ env.GCP_SERVICE_API_3 }}
        gcloud services enable ${{ env.GCP_SERVICE_API_4 }}

Bereitstellungs-Buckets

deploy-buckets:
    needs: [enable-services]
    runs-on: ubuntu-22.04
    timeout-minutes: 10

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker

    # Step to Create GCP Bucket 
    - name: Create Google Cloud Storage - datalake
      run: |-
        if ! gsutil ls -p ${{ secrets.PROJECT_ID }} gs://${{ secrets.BUCKET_DATALAKE }} &> /dev/null; \
          then \
            gcloud storage buckets create gs://${{ secrets.BUCKET_DATALAKE }} --default-storage-class=nearline --location=${{ env.REGION }}
          else
            echo "Cloud Storage : gs://${{ secrets.BUCKET_DATALAKE }}  already exists" ! 
          fi


    # Step to Upload the file to GCP Bucket - transient files
    - name: Upload transient files to Google Cloud Storage
      run: |-
        TARGET=${{ env.INPUT_FOLDER }}
        BUCKET_PATH=${{ secrets.BUCKET_DATALAKE }}/${{ env.INPUT_FOLDER }}    
        gsutil cp -r $TARGET gs://${BUCKET_PATH}

Deploy-Cloud-Funktion

deploy-cloud-function:
    needs: [enable-services, deploy-buckets]
    runs-on: ubuntu-22.04
    steps:
    - uses: actions/checkout@v2

    # Step to Authenticate with GCP
    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Configure  Cloud SDK
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker

    - name: Set up python 3.10
      uses: actions/setup-python@v2
      with:
        python-version: 3.10.12
    #cloud_function_scripts/csv_to_parquet
    - name: Create cloud function - ${{ env.CLOUD_FUNCTION_1_NAME }}
      run: |-
        cd ${{ env.FUNCTION_SCRIPTS }}/${{ env.CLOUD_FUNCTION_1_NAME }}
        gcloud functions deploy ${{ env.CLOUD_FUNCTION_1_NAME }} \
        --gen2 \
        --cpu=${{ env.FUNCTION_CPU  }} \
        --memory=${{ env.FUNCTION_MEMORY  }} \
        --runtime ${{ env.PYTHON_FUNCTION_RUNTIME }} \
        --trigger-http \
        --region ${{ env.REGION }} \
        --entry-point ${{ env.CLOUD_FUNCTION_1_NAME }}

    - name: Create cloud function - ${{ env.CLOUD_FUNCTION_2_NAME }}
      run: |-
        cd ${{ env.FUNCTION_SCRIPTS }}/${{ env.CLOUD_FUNCTION_2_NAME }}
        gcloud functions deploy ${{ env.CLOUD_FUNCTION_2_NAME }} \
        --gen2 \
        --cpu=${{ env.FUNCTION_CPU  }} \
        --memory=${{ env.FUNCTION_MEMORY  }} \
        --runtime ${{ env.PYTHON_FUNCTION_RUNTIME }} \
        --trigger-http \
        --region ${{ env.REGION }} \
        --entry-point ${{ env.CLOUD_FUNCTION_2_NAME }}

    - name: Create cloud function - ${{ env.CLOUD_FUNCTION_3_NAME }}
      run: |-
        cd ${{ env.FUNCTION_SCRIPTS }}/${{ env.CLOUD_FUNCTION_3_NAME }}
        gcloud functions deploy ${{ env.CLOUD_FUNCTION_3_NAME }} \
        --gen2 \
        --cpu=${{ env.FUNCTION_CPU  }} \
        --memory=${{ env.FUNCTION_MEMORY  }} \
        --runtime ${{ env.PYTHON_FUNCTION_RUNTIME }} \
        --trigger-http \
        --region ${{ env.REGION }} \
        --entry-point ${{ env.CLOUD_FUNCTION_3_NAME }}

Deploy-Composer-Service-Konto

deploy-composer-service-account:
    needs: [enable-services, deploy-buckets, deploy-cloud-function ]
    runs-on: ubuntu-22.04
    timeout-minutes: 10

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker


    - name: Create service account
      run: |-

        if ! gcloud iam service-accounts list | grep -i ${{ env.SERVICE_ACCOUNT_NAME}} &> /dev/null; \
          then \
            gcloud iam service-accounts create ${{ env.SERVICE_ACCOUNT_NAME }} \
            --display-name=${{ env.SERVICE_ACCOUNT_DESCRIPTION }}
          fi
    - name: Add permissions to service account
      run: |-
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/composer.user"

        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/storage.objectAdmin"

        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/cloudfunctions.invoker"

        # Permissão para criar e gerenciar ambientes Composer
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/composer.admin"

        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/composer.worker"

        # Permissão para criar e configurar instâncias e recursos na VPC
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/compute.networkAdmin"

        # Permissão para interagir com o Cloud Storage, necessário para buckets e logs
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/storage.admin"

        # Permissão para criar e gerenciar recursos no projeto, como buckets e instâncias
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/editor"

        # Permissão para acessar e usar recursos necessários para o IAM
        gcloud projects add-iam-policy-binding ${{secrets.PROJECT_ID}} \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/iam.serviceAccountUser"

        gcloud functions add-iam-policy-binding ${{env.CLOUD_FUNCTION_1_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/cloudfunctions.invoker"

        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_1_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" 

        gcloud functions add-iam-policy-binding ${{env.CLOUD_FUNCTION_2_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/cloudfunctions.invoker"

        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_2_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com"     

        gcloud functions add-iam-policy-binding ${{env.CLOUD_FUNCTION_3_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
          --role="roles/cloudfunctions.invoker"

        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_3_NAME}} \
          --region="${{env.REGION}}" \
          --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com"   

        SERVICE_NAME_1=$(gcloud functions describe ${{ env.CLOUD_FUNCTION_1_NAME }} --region=${{ env.REGION }} --format="value(serviceConfig.service)")
        gcloud run services add-iam-policy-binding $SERVICE_NAME_1 \
        --region="${{env.REGION}}" \
        --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
        --role="roles/run.invoker"

        SERVICE_NAME_2=$(gcloud functions describe ${{ env.CLOUD_FUNCTION_2_NAME }} --region=${{ env.REGION }} --format="value(serviceConfig.service)")
        gcloud run services add-iam-policy-binding $SERVICE_NAME_2 \
        --region="${{env.REGION}}" \
        --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
        --role="roles/run.invoker"

        SERVICE_NAME_3=$(gcloud functions describe ${{ env.CLOUD_FUNCTION_3_NAME }} --region=${{ env.REGION }} --format="value(serviceConfig.service)")
        gcloud run services add-iam-policy-binding $SERVICE_NAME_3 \
        --region="${{env.REGION}}" \
        --member="serviceAccount:${{env.SERVICE_ACCOUNT_NAME}}@${{secrets.PROJECT_ID}}.iam.gserviceaccount.com" \
        --role="roles/run.invoker"


        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_1_NAME}} \
        --region="${{env.REGION}}" \
        --member="allUsers"

        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_2_NAME}} \
        --region="${{env.REGION}}" \
        --member="allUsers"

        gcloud functions add-invoker-policy-binding ${{env.CLOUD_FUNCTION_3_NAME}} \
        --region="${{env.REGION}}" \
        --member="allUsers"

Deploy-BigQuery-Dataset-BigQuery-Tabellen

deploy-bigquery-dataset-bigquery-tables:
    needs: [enable-services, deploy-buckets, deploy-cloud-function, deploy-composer-service-account ]
    runs-on: ubuntu-22.04
    timeout-minutes: 10

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker


    - name: Create Big Query Dataset
      run: |-  
        if ! bq ls --project_id ${{ secrets.PROJECT_ID}} -a | grep -w ${{ env.BIGQUERY_DATASET}} &> /dev/null; \
          then 

            bq --location=${{ env.REGION }} mk \
          --default_table_expiration 0 \
          --dataset ${{ env.BIGQUERY_DATASET }}

          else
            echo "Big Query Dataset : ${{ env.BIGQUERY_DATASET }} already exists" ! 
          fi

    - name: Create Big Query table
      run: |-
        TABLE_NAME_CUSTOMER=${{ env.BIGQUERY_DATASET}}.${{ env.BIGQUERY_TABLE_CUSTOMER}}
        c=0
        for table in $(bq ls --max_results 1000 "${{ secrets.PROJECT_ID}}:${{ env.BIGQUERY_DATASET}}" | tail -n +3 | awk '{print }'); do

            # Determine the table type and file extension
            if bq show --format=prettyjson $BIGQUERY_TABLE_CUSTOMER | jq -r '.type' | grep -q -E "TABLE"; then
              echo "Dataset ${{ env.BIGQUERY_DATASET}} already has table named : $table " !
              if [ "$table" == "${{ env.BIGQUERY_TABLE_CUSTOMER}}" ]; then
                echo "Dataset ${{ env.BIGQUERY_DATASET}} already has table named : $table " !
                ((c=c+1))       
              fi                  
            else
                echo "Ignoring $table"            
                continue
            fi
        done
        echo " contador $c "
        if [ $c == 0 ]; then
          echo "Creating table named : $table for Dataset ${{ env.BIGQUERY_DATASET}} " !

          bq mk --table \
          $TABLE_NAME_CUSTOMER \
          ./big_query_schemas/customer_schema.json


        fi

Deploy-Composer-Umgebung

deploy-composer-environment:
    needs: [enable-services, deploy-buckets, deploy-cloud-function, deploy-composer-service-account, deploy-bigquery-dataset-bigquery-tables ]
    runs-on: ubuntu-22.04
    timeout-minutes: 40

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker

    # Create composer environments
    - name: Create composer environments
      run: |-
        if ! gcloud composer environments list --project=${{ secrets.PROJECT_ID }} --locations=${{ env.REGION }} | grep -i ${{ env.COMPOSER_ENV_NAME }} &> /dev/null; then
            gcloud composer environments create ${{ env.COMPOSER_ENV_NAME }} \
                --project ${{ secrets.PROJECT_ID }} \
                --location ${{ env.REGION }} \
                --environment-size ${{ env.COMPOSER_ENV_SIZE }} \
                --image-version ${{ env.COMPOSER_IMAGE_VERSION }} \
                --service-account ${{ env.SERVICE_ACCOUNT_NAME }}@${{ secrets.PROJECT_ID }}.iam.gserviceaccount.com
        else
            echo "Composer environment ${{ env.COMPOSER_ENV_NAME }} already exists!"
        fi

    # Create composer environments
    - name: Create composer variable PROJECT_ID 
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION}} variables \
        -- set PROJECT_ID ${{ secrets.PROJECT_ID }}

    - name: Create composer variable REGION
      run: |-  
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
          --location ${{ env.REGION }} variables \
          -- set REGION ${{ env.REGION }}

    - name: Create composer variable CLOUD_FUNCTION_1_NAME
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }}\
          --location ${{ env.REGION }} variables \
          -- set CLOUD_FUNCTION_1_NAME ${{ env.CLOUD_FUNCTION_1_NAME }}

    - name: Create composer variable CLOUD_FUNCTION_2_NAME
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set CLOUD_FUNCTION_2_NAME ${{ env.CLOUD_FUNCTION_2_NAME }}

    - name: Create composer variable CLOUD_FUNCTION_3_NAME
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set CLOUD_FUNCTION_3_NAME ${{ env.CLOUD_FUNCTION_3_NAME }}

    - name: Create composer variable BUCKET_DATALAKE
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION}} variables \
        -- set BUCKET_NAME ${{ secrets.BUCKET_DATALAKE }}

    - name: Create composer variable TRANSIENT_FILE_PATH
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set TRANSIENT_FILE_PATH ${{ env.TRANSIENT_FILE_PATH }}

    - name: Create composer variable BRONZE_PATH 
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set BRONZE_PATH ${{ env.BRONZE_PATH }}

    - name: Create composer variable SILVER_PATH
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set SILVER_PATH ${{ env.SILVER_PATH }}

    - name: Create composer variable REGION_PROJECT_ID
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set REGION_PROJECT_ID "${{ env.REGION }}-${{ secrets.PROJECT_ID }}"


    - name: Create composer variable BIGQUERY_DATASET
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set BIGQUERY_DATASET "${{ env.BIGQUERY_DATASET }}"

    - name: Create composer variable BIGQUERY_TABLE_CUSTOMER
      run: |-
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} variables \
        -- set BIGQUERY_TABLE_CUSTOMER "${{ env.BIGQUERY_TABLE_CUSTOMER }}"

Deploy-Composer-http-Verbindung

deploy-composer-http-connection:
    needs: [enable-services, deploy-buckets, deploy-cloud-function, deploy-composer-service-account, deploy-bigquery-dataset-bigquery-tables, deploy-composer-environment ]
    runs-on: ubuntu-22.04

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    # Step to Configure Docker to use the gcloud command-line tool as a credential helper
    - name: Configure Docker
      run: |-
        gcloud auth configure-docker

    - name: Create composer http connection HTTP_CONNECTION
      run: |-
        HOST="https://${{ env.REGION }}-${{ secrets.PROJECT_ID }}.cloudfunctions.net"
        gcloud composer environments run ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} connections \
        -- add ${{ env.HTTP_CONNECTION }} \
        --conn-type ${{ env.CONNECTION_TYPE  }} \
        --conn-host $HOST

Bereitstellungstage

deploy-dags:
    needs: [enable-services, deploy-buckets, deploy-cloud-function, deploy-composer-service-account, deploy-bigquery-dataset-bigquery-tables, deploy-composer-environment, deploy-composer-http-connection ]
    runs-on: ubuntu-22.04

    steps:
    - name: Checkout
      uses: actions/checkout@v4

    - name: Authorize GCP
      uses: 'google-github-actions/auth@v2'
      with:
        credentials_json:  ${{ secrets.GCP_DEVOPS_SA_KEY }}

    # Step to Authenticate with GCP
    - name: Set up Cloud SDK
      uses: google-github-actions/setup-gcloud@v2
      with:
        version: '>= 363.0.0'
        project_id: ${{ secrets.PROJECT_ID }}

    - name: Get Composer bucket name and Deploy DAG to Composer
      run: |-
        COMPOSER_BUCKET=$(gcloud composer environments describe ${{ env.COMPOSER_ENV_NAME }} \
        --location ${{ env.REGION }} \
        --format="value(config.dagGcsPrefix)")
        gsutil -m cp -r ./dags/* $COMPOSER_BUCKET/dags/

Nach der Bereitstellung erstellte Ressourcen

Nach Abschluss des Bereitstellungsprozesses stehen die folgenden Ressourcen zur Verfügung:

1. Cloud-Speicher-Buckets: Organisiert in Bronze-, Silber- und Goldschichten.
2. Cloud-Funktionen:: Verantwortlich für die Verarbeitung von Daten über die drei Ebenen hinweg.
3. Dienstkonto für Composer:: Mit entsprechenden Berechtigungen zum Aufrufen von Cloud-Funktionen.
4. BigQuery-Datensatz und -Tabellen: Ein DataFrame, der zum Speichern verarbeiteter Daten erstellt wurde.
5. Google Composer-Umgebung: Orchestriert die Cloud-Funktionen mit täglichen Ausführungen.
6. Composer DAG: Der DAG verwaltet den Workflow, der Cloud Functions aufruft und Daten verarbeitet.

Abschluss

Diese Lösung zeigt, wie Sie Google Cloud Functions, Composer und BigQuery nutzen können, um eine robuste dreistufige Datenverarbeitungspipeline zu erstellen. Die Automatisierung mithilfe von GitHub Actions gewährleistet einen reibungslosen, reproduzierbaren Bereitstellungsprozess und erleichtert die Verwaltung cloudbasierter Datenpipelines im großen Maßstab.

Referenzen

1. Google Cloud Platform-Dokumentation: https://cloud.google.com/docs
2. GitHub-Aktionsdokumentation:: https://docs.github.com/en/actions
3. Google Composer-Dokumentation:: https://cloud.google.com/composer/docs
4. Cloud Functions-Dokumentation: https://cloud.google.com/functions/docs
5. GitHub Repo: https://github.com/jader-lima/gcp-cloud-functions-to-bigquery

Das obige ist der detaillierte Inhalt vonVerwendung von Google Cloud-Funktionen für die dreistufige Datenverarbeitung mit Google Composer und automatisierten Bereitstellungen über GitHub-Aktionen. Für weitere Informationen folgen Sie bitte anderen verwandten Artikeln auf der PHP chinesischen Website!