databricks-deploy-integration（databricks 部署集成）

Databricks Deploy Integration

Overview

Deploy Databricks jobs and pipelines using Databricks Asset Bundles (DABs). Asset Bundles provide infrastructure-as-code for deploying jobs, notebooks, DLT pipelines, and ML models across workspaces with proper environment isolation and CI/CD integration.

Prerequisites

• Databricks CLI v0.200+ installed (databricks command)
• Workspace access with appropriate permissions
• Service principal for automated deployments
• databricks.yml bundle configuration

Instructions

Step 1: Initialize Asset Bundle

# Create new bundle from template
databricks bundle init

# Or manually create databricks.yml

# databricks.yml
bundle:
  name: etl-pipeline

workspace:
  host: https://myworkspace.cloud.databricks.com

resources:
  jobs:
    daily_etl:
      name: daily-etl-${bundle.environment}
      schedule:
        quartz_cron_expression: "0 0 6 * * ?"
        timezone_id: "America/New_York"
      tasks:
        - task_key: extract
          notebook_task:
            notebook_path: ./src/extract.py
          new_cluster:
            spark_version: "14.3.x-scala2.12"
            node_type_id: "i3.xlarge"
            num_workers: 2
        - task_key: transform
          depends_on:
            - task_key: extract
          notebook_task:
            notebook_path: ./src/transform.py

environments:
  development:
    default: true
    workspace:
      host: https://dev.cloud.databricks.com
  staging:
    workspace:
      host: https://staging.cloud.databricks.com
  production:
    workspace:
      host: https://prod.cloud.databricks.com

Step 2: Deploy to Environment

# Validate bundle configuration
databricks bundle validate -e production

# Deploy resources (create/update jobs, notebooks)
databricks bundle deploy -e staging

# Run a specific job
databricks bundle run daily_etl -e staging

# Destroy resources in an environment
databricks bundle destroy -e development

Step 3: CI/CD Pipeline

# .github/workflows/deploy.yml
name: Deploy Databricks Bundle
on:
  push:
    branches: [main]

jobs:
  deploy-staging:
    runs-on: ubuntu-latest
    environment: staging
    steps:
      - uses: actions/checkout@v4
      - uses: databricks/setup-cli@main
      - run: databricks bundle validate -e staging
        env:
          DATABRICKS_HOST: ${{ secrets.DATABRICKS_HOST }}
          DATABRICKS_TOKEN: ${{ secrets.DATABRICKS_TOKEN }}
      - run: databricks bundle deploy -e staging
        env:
          DATABRICKS_HOST: ${{ secrets.DATABRICKS_HOST }}
          DATABRICKS_TOKEN: ${{ secrets.DATABRICKS_TOKEN }}

  deploy-production:
    needs: deploy-staging
    runs-on: ubuntu-latest
    environment: production
    steps:
      - uses: actions/checkout@v4
      - uses: databricks/setup-cli@main
      - run: databricks bundle deploy -e production
        env:
          DATABRICKS_HOST: ${{ secrets.DATABRICKS_PROD_HOST }}
          DATABRICKS_TOKEN: ${{ secrets.DATABRICKS_PROD_TOKEN }}

Step 4: Verify Deployment

# List deployed jobs
databricks jobs list --output json | jq '.[] | select(.settings.name | contains("etl"))'

# Check recent runs
databricks runs list --job-id $JOB_ID --limit 5

# Get run output
databricks runs get-output --run-id $RUN_ID

Error Handling

Examples

Basic usage: Apply databricks deploy integration to a standard project setup with default configuration options.

Advanced scenario: Customize databricks deploy integration for production environments with multiple constraints and team-specific requirements.

Resources

• Databricks Asset Bundles
• Databricks CLI
• CI/CD with Bundles

Next Steps

For multi-environment setup, see databricks-multi-env-setup.

Output

• Configuration files or code changes applied to the project
• Validation report confirming correct implementation
• Summary of changes made and their rationale