A practical look at resolver-level caching in Graphene-Django: when it helps, what it hides, and why it should be treated as a tactical workaround rather than good API design.