Nowadays, we are witnessing an explosion in the number of Internet of Things (IoT) devices spread all around the world along with an increasing demand of efficient pervasive distributed IoT as a Service (IoTaaS). In this context, due to the limited hardware capabilities of IoT devices, the latest trend to arrange IoTaaS is to integrate containerized microservices spread over the Cloud and the Edge of the network. However, the management IoTaaS instances is not trivial due to their underlying complexity and the huge amount of data that have to be considered. In fact, the distributed microservices composing each IoTaaS instance must be continuously monitored in order to guarantee a given Service Level Agreement (SLA). In this paper, we discuss how a NoSQL Graph DataBase Management System (DBMS) can help IoT Cloud providers to monitor and optimize IoTaaS instances. In our experimentation, by considering an IoTaaS smart meeting case study, we compare one of the major NoSQL Graph DBMS, i.e., Neo4j with one of the major Relational DBMS, i.e, Oracle, proving that the first one well scale up than the second one when the number of relations increases.