%20--%3e%3cstyle%20type='text/css'%3e.cls-0%20{fill:%23FF9900;}%3c/style%3e%3cpath%20class='cls-0'%20d='m117.3%2084h-14.1l39.2-79h-24.8l-20.3%2041.7-20.5-41.7h-3.8l-20.3%2041.7-20.8-41.7h-24.3l41%2083.3h8.2l18.2-37.7%2018.3%2037.8h23.7v2.8h10.7v-10.5h-10.6v3.3h0.2zm3.3-74.7h15.2l-28.9%2057.1-7.2-14.9%2020.6-42.2h0.3zm-69.3%2074.9-37-75h15.1l20.8%2042.4-5.8%2011%204.1%201.6%2016.9-33.9%207.2%2015.7-18.4%2038.1-2.9%200.1zm44.5-0.1-28.2-58.9%207.4-14.8%2029.6%2060.8-6.1%2012.9c0%200.1-2.7%200-2.7%200zm27.9%200.6v2.8h-3.1v-2.8h3.1z'/%3e%3c/svg%3e)
847 workloads to AWS in 14 weeks, without a production incident
847 workloads in 14 weeks is roughly sixty a week, sustained. At that rate the constraint is never any individual workload — it is knowing what depends on what, and being certain enough of the answer to move things in the right order.
- 847 workloads migrated to AWS
- Delivered in 14 weeks
- Zero production incidents during the programme
What made it hard
847 workloads in 14 weeks is roughly sixty a week, sustained. At that rate the constraint is never any individual workload — it is knowing what depends on what, and being certain enough of the answer to move things in the right order.
Retail also removes the easy option. There is no quiet quarter to migrate in, and a failed cutover is not an internal inconvenience; it is revenue not taken. The programme had to assume that any wave could go wrong and still not take trade down with it.
How it was done
Discovery before scheduling
Dependencies were mapped before any wave was planned. Migration order that follows the org chart rather than the dependency graph is the usual reason a cutover fails at 2am.
Waves sized to be reversible
Each wave was small enough to roll back inside its own window. Reversibility is what makes a fast cadence safe; without it, speed just increases the blast radius.
Cutovers rehearsed, not improvised
Each cutover was run through before the real one, so the steps and the failure points were known in advance rather than discovered live.
Landing zone first
Accounts, identity, network boundaries and guardrails were in place before workloads arrived, so controls did not have to be retrofitted across 847 of them later.
Where it landed
All 847 workloads moved inside the 14-week window, with no production incidents recorded against the migration.
The absence of incidents is the number worth reading. Migrations of this size usually buy their speed with a few bad nights; this one did not.
Bring us the version of this you are facing
A 45-minute call with the engineer who would run it. We will tell you what we would look at first, and whether it is a project at all.
Talk to an Engineer