Every team that splits a monolith into services expects to get microservices. What they often get instead is a distributed monolith — a system with all the operational overhead of separate services and none of the independence that was supposed to justify it.
The tell is always the same: deploys are still coupled, a schema change in one service still requires coordinated releases across three others, and nobody can tell you where a request actually terminates without tracing it live. Splitting a codebase along file boundaries is not the same as splitting it along ownership boundaries.
The fix isn’t more services — it’s fewer, better-drawn ones. A boundary is only real if the team on either side of it can deploy, fail, and scale independently. If they can’t, you haven’t decomposed the monolith. You’ve just given it worse latency.