I do in fact think it's pretty unlikely that any performance degradation you observed was directly caused by microservices, as opposed to changes that directly affected the frontend.