部署自动回滚咨询：失败部署能否执行自动回滚及相关文档查询

Great question! Yes, automatic rollbacks for failed deployments are absolutely supported across most modern deployment tools and platforms—let me break this down for you, including where to find the official docs you’re looking for:

How to Implement Automatic Rollbacks #

The exact setup depends on your tooling, but here are the most common approaches:

• Orchestration Platforms (e.g., Kubernetes)
  Kubernetes natively supports automatic rollbacks tied to health checks. Configure readiness and liveness probes for your containers, and set deployment strategy constraints like maxUnavailable and maxSurge. If the new version fails health checks during rollout, Kubernetes will automatically revert to the last successful revision.

  Example Deployment snippet with auto-rollback triggers:

  apiVersion: apps/v1
  kind: Deployment
  metadata:
    name: my-app-deployment
  spec:
    replicas: 3
    strategy:
      rollingUpdate:
        maxSurge: 1
        maxUnavailable: 0
    selector:
      matchLabels:
        app: my-app
    template:
      metadata:
        labels:
          app: my-app
      spec:
        containers:
        - name: my-app-container
          image: my-app:v2
          readinessProbe:
            httpGet:
              path: /health
              port: 8080
            initialDelaySeconds: 5
            periodSeconds: 10
          livenessProbe:
            httpGet:
              path: /health
              port: 8080
            initialDelaySeconds: 15
            periodSeconds: 20
  

• CI/CD Pipelines (e.g., GitHub Actions, GitLab CI)
  You can add conditional steps in your pipeline to trigger a rollback if deployment validation fails. For example, in GitHub Actions, you might run a health check after deployment, and if it fails, execute a rollback command like kubectl rollout undo deployment/my-app-deployment.

Official Documentation References #

Since you only found manual rollback docs before, here’s where to locate auto-rollback details for major platforms:

• Kubernetes: Check the official docs sections on Deployment Rollbacks and Container Probes—the auto-rollback behavior is explicitly tied to probe failures during rollouts.
• AWS Elastic Beanstalk: Look for the Rollback on Failure setting in environment deployment configurations; the docs explain how enabling this triggers automatic rollbacks on deployment health check failures.
• Azure App Service: Use deployment slots with auto-swap enabled, paired with health checks. If the swap fails health validation, the platform automatically swaps back to the previous stable slot—this is covered in the Deployment Slots documentation.

Key Tips #

• Health checks are non-negotiable: Auto-rollback relies on your tool being able to detect a failed deployment. Ensure your readiness/liveness probes accurately reflect your app’s health.
• You can still trigger manual rollbacks anytime (e.g., kubectl rollout undo for Kubernetes) if needed, but auto-rollback eliminates manual intervention for most failure scenarios.

内容的提问来源于stack exchange，提问作者Pacolotero