Cutover runbook best practices

Plan on scheduling a minimum of two walkthrough meetings to ensure all the necessary details have been captured.

Invite all identified task owners to the walkthrough meetings so they are aware of the scope of the tasks and when exactly they need to perform their tasks. Tasks might be scheduled for nighttime, early morning, and so forth.

Walkthrough meetings offer an additional opportunity to include stakeholders (for example, enterprise integration testers) of the applications that might be directly or indirectly impacted as a result of the migration. Participation of members from these satellite teams might be needed for final sign-off.

Specifically identify the individual people responsible for performing the tasks, rather than the entire teams, to establish explicit ownership.

Ideally, schedule the final walk-through of the cutover runbook for 2-3 days before the actual cutover. This is to ensure that any last-minute blockers are identified and raised in front of all stakeholders.

Ensure that the link to the live document is shared with all stakeholders before the cutover so that they can reference the cutover runbook during the migration window.

As tasks are completed, update the cutover runbook with the progress. This provides stakeholders with an up-to-date status on the state of activities.

To minimize the downtime and impact to production users, identify and group tasks that can be performed in parallel Examples of these tasks include taking the backups of the source databases and launching new Amazon Relational Database Service (Amazon RDS) instances in the target environment.

Identify the outage window within which the migration needs to be successfully completed. Align on a rollback plan in case the migration exceeds the outage window.

To benefit future migrations, it’s essential to capture in the document any observations and troubleshooting that might occur during the migration.