Skip to main content
OpsRoadmaps

Checklists / cloud-migration

Cloud Migration Checklist

Move workloads to the cloud without surprises: scope, cost, security, egress, and an exit story.

9 steps · 9 required · last reviewed (next review 2027-03-19) · updated 2026-09-21

Related focus areas:CloudOpsFinOpsInfraOps

Roadmaps covering the same ground:AWS Security ImprovementCloud Cost OptimizationObservability MaturityProduction-Ready KubernetesResilience and DR Improvement

Sign in to run this checklist

Cloud migrations fail in the planning phase and only reveal it during execution. This checklist runs at planning time — every unchecked box is a decision being made by default.

The two boxes teams regret skipping: egress costs and the exit plan. Cloud providers make walking in cheap and walking out expensive; know both numbers before you start.

  1. Workloads are inventoried with dependencies mapped (including the ones nobody documented)

    required

    Undiscovered dependencies are why 'lift and shift' projects stall at 80% — the forgotten nightly job that reads the old database.

  2. Target-state architecture is decided (rehost / replatform / refactor per workload)

    required

    Lifting everything unchanged inherits every problem you have; refactoring everything makes the migration take years. Decide per workload, in writing.

  3. Cost model includes egress, storage tiers, managed-service premiums, and idle baseline

    required

    Compute is cheap; the bill arrives from data leaving, data sitting, and services running 24/7 for 5% utilization.

  4. Security model is designed (IAM structure, network segmentation, encryption, logging)

    required

    Retro-fitting security after workloads land means an unflagged gap period exactly when everything is in flux.

  5. Data migration plan covers initial sync, incremental sync, and the cutover window

    required

    Databases moved in one big-bang copy mean a long outage and a rollback cliff; sync-and-cut-over means minutes.

  6. Cutover and rollback criteria are defined per workload (what success looks like, when to abort)

    required

    'It feels done' is not a cutover criterion; latency, error rate, and data-consistency numbers are.

  7. Compliance and data-residency requirements are verified in the target region

    required

    Some data legally cannot cross some borders; discovering this post-migration is a project-restart.

  8. Observability is live in the cloud BEFORE workloads arrive

    required

    Migrating blind means the first week in the cloud is also the first week without metrics — precisely when everything is unfamiliar.

  9. Exit plan exists: data export paths, portability choices, and their cost

    required

    Multi-year commitments with no exit story convert a bad pricing change next year into a hostage situation.

Run it against your real gaps

A checklist is most useful when you know which controls you are actually missing. A free OpsRoadmaps assessment scores your controls with explainable findings and builds a roadmap from the results.

Assess my stack

Checklists document self-reported practice. Items were last reviewed 2026-09-19 — adapt them to your own environment.