# Review the cutover's last reversible checkpoint

A rollback instruction needs a data boundary and enough time to execute. Identify exactly when the old environment stops being a safe destination.

By Cobnex editorial. Published 2026-09-10. Updated 2026-09-11.

## Mark the change in write authority

Locate the step that allows the target to accept business changes. Before that step, source recovery may be straightforward. Afterwards, the target can contain records the source lacks.

Review the plan for those new records. It may use a tested return path, controlled reconciliation or forward repair. The plan must not assume they disappear because traffic is routed elsewhere.

Include background and partner writes in this analysis. The first target write may arrive from a queue before the public route is opened.

## Check the evidence for data readiness

Ask how the final source checkpoint is identified and how its application at the target is verified. Low replication lag is useful but does not replace a defined completion condition.

Review excluded tables, files and external state. A migration tool's success status applies to its configured scope, which may not cover everything the application needs.

Inspect reconciliation examples for active business records, not only static reference data. The most recent work is often the most important during reopening.

## Calculate the latest decision time

Work backwards from the end of the permitted interruption. Reserve time for the measured recovery procedure and its verification.

If the team reaches that decision time without passing the required checks, the runbook should specify the action. Waiting for one more uncertain test can consume the recovery option.

Use rehearsal timings with a reasonable allowance for uncertainty. A theoretical command duration does not include coordination, data checks or user-path validation.

## Confirm who can act

Verify the decision owner's availability and the technical access needed for each recovery step. A procedure requiring an unavailable provider or administrator is not executable within the window.

Check communication paths for users and support teams, including the expected state during a pause. Avoid declaring service restored before the business workflow has been verified.

Approve the cutover when the checkpoint, evidence and recovery action are concrete. A long runbook without these decisions can still leave the team improvising at the moment the system becomes hardest to reverse.

## Sources

- [AWS: pre-cutover planning](https://docs.aws.amazon.com/prescriptive-guidance/latest/best-practices-migration-cutover/pre-cutover-stage.html)