that will happen even if the two "root" API requests hit different SpiceDB nodes, since SpiceDB is designed to cache on specific nodes within the cluster for specific subproblems