Discover what must move and what must remain.
Inventory websites, applications, databases, files, scheduled tasks, certificates and mail services. Record DNS ownership, current access, software versions and licences. Map integrations and source restrictions that may depend on an IP address or a particular hosting location.
Name the business owner and technical contacts for each component. Agree the migration scope and success criteria before scheduling the work. Retain a recoverable source copy and establish who can authorise changes to the live service.
Validate the destination with realistic checks.
Build a test copy and check application compatibility, permissions, storage and network access. Use a controlled way to reach the destination before changing public DNS. Review forms, authentication, checkout, uploads, redirects and background jobs as appropriate.
Keep test activity from creating real customer messages or transactions. Record defects and resolve them before the final window. Compare representative performance and verify monitoring and backups, so the destination is ready to operate rather than merely ready to receive files.
Protect changing data and communications.
Choose a supported method for transferring databases in a consistent state. Account for writes that occur during the move and decide when they must stop or be synchronised. Validate records and application behaviour after the final transfer.
Business email needs its own plan where it is in scope. Review mailboxes, aliases, delivery routes and authentication records. Keep website hosting changes separate from mail routing changes unless the migration explicitly includes both, and verify incoming and outgoing messages through approved tests.
Control the switch and its acceptance.
Review DNS records and caching before the window. Reducing a record’s TTL in advance can help shorten later caching, but existing cached responses persist according to their earlier settings. Plan for traffic reaching either environment during the transition.
Define rollback triggers, a decision owner and how new writes would be reconciled if the source must resume service. After cutover, check user journeys, logs, scheduled jobs and integrations. Retire the old environment only after agreed acceptance and retention steps are complete.
