Cloud Migration Checklist: Plan, Move and Optimise

Before You Start

A cloud migration succeeds or fails in planning. Moving servers is the easy part; understanding dependencies, data and people is the hard part. Use this checklist whether you are moving to AWS, Microsoft Azure or Google Cloud.

Phase 1: Assess

• Inventory every application, database, scheduled job and integration.
• Map dependencies between systems, including the undocumented ones.
• Record current performance, availability and monthly cost as a baseline.
• Identify compliance requirements such as data residency, HIPAA or PCI DSS.
• Agree the business goal: lower cost, better resilience, faster releases or retiring a data centre.

Phase 2: Choose a Strategy for Each Application

The widely used "6 Rs" framework:

1. Rehost: lift and shift with minimal change. Fastest, but keeps old inefficiencies.
2. Replatform: small changes such as moving to a managed database.
3. Refactor: redesign for cloud-native services. Most effort, most long-term benefit.
4. Repurchase: replace with a SaaS product.
5. Retire: switch off what nobody uses.
6. Retain: keep on-premises for now.

For tightly coupled legacy systems, an incremental approach such as the Strangler Fig pattern reduces risk.

Phase 3: Build the Landing Zone

• Account or subscription structure, separated by environment.
• Identity and access management with least privilege and single sign-on.
• Network design, encryption defaults and logging.
• Infrastructure as code (for example Terraform) so environments are reproducible.
• Budgets and cost alerts from day one.

Phase 4: Migrate

• Start with a low-risk application to prove the process.
• Replicate data continuously before cutover to keep downtime short.
• Automate deployments with a CI/CD pipeline. See our DevOps and CI/CD service.
• Write a cutover runbook with a rollback plan and named owners for each step.

Phase 5: Validate

• Compare performance against the baseline.
• Run functional, security and failover tests.
• Confirm monitoring, alerting and backups work as expected.
• Get sign-off from business owners.

Phase 6: Optimise

• Right-size instances based on real usage.
• Use autoscaling and managed services to reduce idle capacity.
• Consider commitment discounts once usage is stable.
• Review costs monthly and tag every resource by owner and environment.

Common Mistakes

• Migrating without a dependency map.
• Underestimating data transfer time for large databases.
• Leaving old infrastructure running and paying twice.
• Treating security as a final step instead of a design input.

Our cloud architecture and migration team plans and executes migrations with zero-downtime cutover strategies.

Frequently Asked Questions

How long does a cloud migration take?

A single application can move in weeks. A full data-centre exit usually takes several months and is done in waves.

Will moving to the cloud reduce my costs?

It can, but only with right-sizing, autoscaling and active cost management. A pure lift and shift sometimes costs more at first.

Which cloud provider should I choose?

Choose based on your team's skills, the managed services you need, compliance requirements and existing licences. All three major providers are capable.

How do I avoid downtime during migration?

Replicate data continuously, test the cutover in rehearsal, and switch traffic gradually with a tested rollback plan.