The safest legacy rewrite is often no rewrite

Unknown behavior is still behavior

A mature service contains decisions that no document remembers. Some are obsolete. Others are the reason a customer workflow still functions. A clean rewrite treats both groups as equally disposable.

Before choosing a replacement, make the system observable enough to identify its actual failure modes, consumers, dependencies, and recovery behavior.

Reduce fear in descending order

Add behavioral tests around the paths that matter. Remove dead integrations. Isolate unstable dependencies. Write down recovery. These steps often solve the operational problem that made a rewrite feel urgent.

If replacement is still justified, the same evidence provides a safer migration boundary.