How I Review a Backend Before Recommending a Rewrite

A rewrite is expensive, emotionally attractive, and often unnecessary. Before recommending one, I inspect the runtime behavior, data model, deployment path, authentication boundary, and the parts of the system that create the most operational pain.

The first question is not whether the code is beautiful. The first question is whether the system can still support the business with targeted intervention. I look for the smallest set of changes that improves reliability, security, and delivery speed.

Only after that evidence is clear do I recommend larger reconstruction.