article --summary
The summary
This AWS customer story explains how British Airways is working to modernize a large legacy Java estate after finishing a move of its data center workloads to AWS. The team chose AWS Transform custom, a tool that applies agentic AI and automation to big migration jobs, to handle more than 250 business services spread across roughly 750 code repositories.
The reader will see the 20-step process used for each service. It starts with deterministic steps that pull production artifacts, study live traffic and set up baseline tests, followed by agentic transformation that upgrades the code to Java 21 and moves it onto Amazon ECS with Fargate and Graviton. The final phase repeats the regression tests, compares results to the baseline and adds human code reviews and security approvals before go-live.
The article also covers the reasons behind the project, including the engineering time spent maintaining the old platform, and the outcomes BA reports so far. Leaders quoted in the piece say the early results are encouraging and that the lessons are being carried to other parts of the platform.
Editorial note: This is a summary of an article first published by AWS DevOps Blog. The full piece is on their site.