Microservice generation system 102 can provide technical improvements to systems, devices, components, operational steps, and/or processing steps associated with the various technologies identified above. For example, microservice generation system 102 can train a model to learn cluster assignments and/or graph embeddings of classes in a monolithic application based on runtime traces of business function test cases that can be executed using the monolithic application; and/or can further employ the model to generate clusters of the classes based on the cluster assignments and/or graph embeddings to identify one or more microservices of the monolithic application. Execution of such business function test cases using the monolithic application can yield runtime traces corresponding to each of such business function test cases. Microservice generation system 102 can utilize such runtime traces that correspond to respective business function test cases to generate causal graphs comprising causal sequences that also correspond to such respective business function test cases. Such causal graphs and/or causal sequences that correspond to respective business function test cases provide microservice generation system 102 with first order temporal dependencies and/or high order temporal dependencies of the monolithic application and/or the cluster assignments of the classes in the monolithic application, where such dependencies correspond to the respective business function test cases. These first order temporal dependencies and/or high order temporal dependencies provide microservice generation system 102 with the various business functionalities of the monolithic application such as, for instance, the order and frequency of how classes and function calls of the monolithic application interact. The first order temporal dependencies and/or the high order temporal dependencies can further enable microservice generation system 102 to associate business context with each trace to provide improved functional grouping (e.g., functional clustering) of the classes in the monolithic application. Such improved functional grouping of the classes in the monolithic application can enable microservice generation system 102 to more accurately identify one or more microservices of the monolithic application, thereby facilitating improved application modernization of the monolithic application by microservice generation system 102.