After a failed deployment, what is a common rollback procedure?

After a failed deployment, what is a common rollback procedure?

• A. Ignore and monitor
• B. Continue with the failing version
• C. Fix in place and redeploy
• D. Revert to the previous known-good version (or perform a blue-green rollback)

Answer: (D)

When a deployment fails, the goal is to restore service quickly by reverting to a version that is known to work. The common rollback approach is to revert to the previous known-good release, often using a blue-green rollback. In a blue-green setup you have two production environments: one running the last stable release and one with the new, possibly faulty release. If problems arise, you switch traffic back to the stable environment, restoring normal service with minimal downtime. This provides a safe, tested state to fall back to and a fast path to recover. Other options don’t offer the same rapid, risk-free recovery. Ignoring the failure leaves users on a broken build; continuing with the failing version prolongs outages and can worsen impact; fixing in place and redeploying can be appropriate, but it’s not the standard rollback method because it assumes the fix is ready and validated before redeploying, whereas rollback prioritizes returning to a known-good state immediately.

When a deployment fails, the goal is to restore service quickly by reverting to a version that is known to work. The common rollback approach is to revert to the previous known-good release, often using a blue-green rollback. In a blue-green setup you have two production environments: one running the last stable release and one with the new, possibly faulty release. If problems arise, you switch traffic back to the stable environment, restoring normal service with minimal downtime. This provides a safe, tested state to fall back to and a fast path to recover.

Other options don’t offer the same rapid, risk-free recovery. Ignoring the failure leaves users on a broken build; continuing with the failing version prolongs outages and can worsen impact; fixing in place and redeploying can be appropriate, but it’s not the standard rollback method because it assumes the fix is ready and validated before redeploying, whereas rollback prioritizes returning to a known-good state immediately.