I rebuild AWS setups clicked together by hand. Or build one from an empty account.

When a rebuild makes sense

• A deploy takes most of an hour, and one person knows how to run it.
• There is no copy of production to test on, so customers find the bugs.
• Nobody can say what happens when a server or a zone goes down.
• The bill keeps growing and nobody can point to the reason.

If there is just one problem, such as a slow or broken deploy, that is usually a few hours of work rather than a rebuild. That is what DevOps by the hour is for.

When AWS makes sense

AWS pays off when traffic swings hard over the day, when you rely on managed services you do not want to run yourself, or when your customers require compliance certifications such as SOC 2. If none of these apply to you, I will say so before the migration.

How I do it

• AWS Organizations with separate accounts per environment and single sign-on through IAM Identity Center.
• Network, permissions and services in Terraform, in your repository.
• A new environment next to the old one, switched over once you approve, with a rollback plan.
• CI/CD with zero-downtime deploys.
• Budgets with alerts, AWS Cost Anomaly Detection and tags from day one.

Example

I built two SaaS platforms on AWS from an empty account. On the first, deploys take 5–6 minutes instead of 45, and the bill is 30% lower than on the original infrastructure its developers had clicked together by hand on ready-made AWS services such as Amplify. The whole story.

I rebuilt a Magento shop the same way, for traffic peaks and failures. The hosting bill went up there, because the goal was scalability and resilience to failure. The whole story.

If the shop is simply slow and does not need to move, start with a review of the shop’s server.

Price

Every rebuild and migration is different, so I quote it after a short call and a look at what runs today. The price is fixed.