Which approach updates production servers and allows rollback by repeating the process?

Which approach updates production servers and allows rollback by repeating the process?

• A. Rolling deployment
• B. Canary deployment
• C. Blue-green deployment
• D. A/B testing

Answer: (A)

Rolling deployment updates production servers in small, controlled batches. You move through the fleet step by step, monitoring each batch as you go. If issues arise, you can halt and re-run the deployment process to revert to the previous version on those servers, effectively rolling back by repeating the same update steps with the older release. This approach minimizes risk and downtime because the change is incremental and reversible. The other methods don’t fit as neatly. Canary deployment also rolls out to a small initial group, but its primary purpose is testing a release on a tiny subset before wider adoption, not the straightforward stepwise update of the entire production set. Blue-green deployment relies on switching traffic between two identical environments, so rollback means flipping traffic back rather than repeatedly updating the same servers in place. A/B testing focuses on comparing variants for performance or user response, not on deployment and controlled rollback of production servers.

Rolling deployment updates production servers in small, controlled batches. You move through the fleet step by step, monitoring each batch as you go. If issues arise, you can halt and re-run the deployment process to revert to the previous version on those servers, effectively rolling back by repeating the same update steps with the older release. This approach minimizes risk and downtime because the change is incremental and reversible.

The other methods don’t fit as neatly. Canary deployment also rolls out to a small initial group, but its primary purpose is testing a release on a tiny subset before wider adoption, not the straightforward stepwise update of the entire production set. Blue-green deployment relies on switching traffic between two identical environments, so rollback means flipping traffic back rather than repeatedly updating the same servers in place. A/B testing focuses on comparing variants for performance or user response, not on deployment and controlled rollback of production servers.