The Missing Time in Microservice Granularity

Choosing microservice granularity is a consequential architectural decision, yet research has focused primarily on how to decompose systems rather than when service boundaries should be reassessed. This systematic mapping examines 52 documents and includes 31 primary studies. It classifies the studies across seven granularity levels (G0–G6), 13 target metrics, and six operational criteria. Business capability microservices (G3) dominate the literature. Splitting and merging services are both reported, but no included study proposes a recurring process for managing granularity, defines a temporal saturation signal, or reports a trigger for reassessment. Only four studies include longitudinal analysis. Structural criteria such as coupling and domain alignment are common, while operational effects such as deployment failure, defects, and monetary cost are rarely measured. The findings point to a need for continuous granularity management supported by temporal metrics, explicit reassessment triggers, and measured operational outcomes.