VM-to-Cloud Cutover Checklist
A cutover checklist for rehosting on-prem VMs to cloud compute: tested replication, right-sized targets, a final incremental sync, DNS repointing, and a soak period that preserves rollback.
When to Use This Checklist
Use this for the cutover step of a rehost (lift-and-shift) migration, when moving on-premises VMs to cloud compute such as EC2, Azure VMs, or Compute Engine, typically with a migration service like AWS MGN, Azure Migrate, or GCP Migrate to Virtual Machines. It assumes assessment and target design are done and focuses on executing the switch with minimal downtime and a clear rollback path.
How to Use This Checklist
Replicate and test before you commit. Boot the replicated VM in an isolated network and validate the application end to end so cutover holds no surprises. Right-size the target to the source's real CPU, memory, and IOPS, not a guess.
During the window, freeze the source, run a final incremental replication to keep RPO small, then repoint DNS and integrations. Keep the source VM intact through a soak period so you can roll back. Only decommission and right-size for cost after the new instance has proven stable.
What Good Looks Like
A clean VM cutover has a tested replication that boots and runs in the cloud, a final incremental sync that keeps data loss near zero, and DNS and integrations repointed within the window. The application passes health, performance, and connectivity checks; monitoring, backups, and agents are confirmed running; and the source VM remains available for rollback through soak. After soak, the source is decommissioned and the instance is right-sized.
Common Pitfalls
The classic mistake is decommissioning the source immediately, leaving no rollback when a problem surfaces hours later. Skipping the isolated test cutover means discovering boot or driver issues during the real window. Oversizing the target wastes money; undersizing degrades performance. Forgetting to reinstall monitoring and backup agents leaves the new instance unprotected. And a stale final replication carries more data loss than expected.
Related Resources
Use the 7 Rs strategy to confirm rehost is the right disposition, and the on-premise-to-cloud migration guides for provider-specific replication tooling. Capacity planning and FinOps practices support the sizing and right-sizing steps.