Rollback

A rollback is the act of returning an application, configuration, or database to an earlier working state after a release causes problems. It is usually the fastest way to stop an incident that started with a deploy.

Why it matters

When errors spike right after a release, the quickest route to recovery is almost always undoing the change, then investigating calmly. Teams that can roll back in minutes treat deployments as routine. Teams that cannot tend to deploy rarely and in large, riskier batches.

Example

A typical decision flow after a deploy:

1. Error rate rises from 0.2 percent to 6 percent two minutes after release.
2. The on-call engineer checks the deploy log, sees the timing matches, and starts the rollback.
3. The previous container image or build is redeployed, for example with kubectl rollout undo deployment/web.
4. Monitoring confirms error rate and response time return to baseline.

Roll back first, ask why second. In blue-green deployments this can be as simple as switching traffic back.

Common mistakes

• Rolling back code but not the database migration it depended on, leaving old code against a new schema. Make migrations backward compatible.
• Never practicing the rollback, so the first attempt happens during an incident.
• Deleting the previous version's artifacts too early.
• "Fixing forward" under pressure when a revert is available and safer.

Related

• Canary release
• Blue-green deployment
• Feature flag
• MTTR

यह पेज सामान्य जानकारी है, जिसे हमारी संपादकीय टीम ने लिखा है और दस्तावेज़ीकरण बदलने पर अपडेट किया जाता है। सर्वर और प्रदाता अलग-अलग होते हैं, इसलिए बदलाव पहले सुरक्षित तरीके से जाँचें। देखें: संपादकीय नीति और अस्वीकरण.