Before you move anything

  • What actually runs today, and what can be retired instead of moved?
  • Which workloads are stateful, and where does their data live?
  • What depends on IP addresses, DNS names, or network locality?
  • Who uses each system, and what downtime can they absorb?

Workloads and data

Move the easiest stateless workload first to prove the path, not the most important one. Data gets its own plan: how it syncs, how long cutover takes, and how you verify nothing was lost. Rehearse the data migration against a copy before doing it for real.

Delivery paths

A migrated workload with no CI/CD is a stranded workload. Plan for the deployment pipeline, secrets, and monitoring to arrive with the workload — not weeks later. The migration is done when the team can ship a change to the new environment, not when the old server is off.

Rollback and readiness

Every cutover needs a written rollback: what triggers it, who decides, and how long the old environment stays warm. Most rollbacks are never used. The ones that are needed are needed badly.