Various embodiments include methods for determining a resource configuration of a cloud service system using a system simulation tool with a model library. This may include: obtaining a model required for each service; obtaining a resource limit setting of each required model; creating a service workflow among the models to obtain a system simulation model, indicating the interaction workflow among the models and the sequence of model execution in a single service; using training data to train a service performance model of each service, wherein the training data comprises test data collected from the cloud service system in a testing environment, and the service performance model defines a resource consumption of a single request running on a service in the system simulation model; performing system simulation with a given resource configuration set and the resource limit settings, to obtain a system performance KPI; and determining a resource configuration with the best KPI.