# Azure Pipelines: Multi-Stage Deployment

Deploy to staging, test, then promote to production with approval.

```yaml title="azure-pipelines.yml"
trigger:
  - main

pool:
  vmImage: ubuntu-latest

stages:
  - stage: DeployStaging
    displayName: "Deploy to Staging"
    jobs:
      - job: Deploy
        steps:
          - task: NodeTool@0
            inputs:
              versionSpec: "20.x"

          - script: npm install

          - script: |
              set -o pipefail
              npx zuplo deploy --api-key $(ZUPLO_API_KEY) --environment staging 2>&1 | tee ./STAGING_STDOUT
              STAGING_URL=$(grep -oP 'Deployed to \K(https://[^ ]+)' ./STAGING_STDOUT)
              echo "##vso[task.setvariable variable=STAGING_URL;isOutput=true]$STAGING_URL"
            name: deployStep
            displayName: "Deploy to staging"

  - stage: TestStaging
    displayName: "Test Staging"
    dependsOn: DeployStaging
    variables:
      STAGING_URL:
        $[
        stageDependencies.DeployStaging.Deploy.outputs['deployStep.STAGING_URL']
        ]
    jobs:
      - job: Test
        steps:
          - task: NodeTool@0
            inputs:
              versionSpec: "20.x"

          - script: npm install

          - script: npx zuplo test --endpoint $(STAGING_URL)
            displayName: "Run tests against staging"

  - stage: DeployProduction
    displayName: "Deploy to Production"
    dependsOn: TestStaging
    jobs:
      - deployment: Deploy
        environment: production
        strategy:
          runOnce:
            deploy:
              steps:
                - checkout: self

                - task: NodeTool@0
                  inputs:
                    versionSpec: "20.x"

                - script: npm install

                - script:
                    npx zuplo deploy --api-key $(ZUPLO_API_KEY) --environment
                    production
                  displayName: "Deploy to production"
```

## Setting Up Approval

Create an environment with approval checks:

1. Go to **Pipelines** > **Environments**
2. Create an environment named `production`
3. Add **Approvals and checks**
4. Add required approvers

The pipeline pauses before production deployment until approved.