Service failure in a service robot is an event that occurs when the delivered service deviates from the correct original service specified by the developers. The cause of failures is due to faults in the robot system, which can be detected based on a model. However, the monitoring task that compares the model and system's behavior is overload. In this study, we propose a multi-layered adaptive monitoring method that complements model-based fault detection. When the target component can be monitored according to their priority adaptively, it results in keeping the efficiency of fault detection, while the overload is reduced.