A financial services company is deploying multiple machine learning models on Amazon SageMaker to provide real-time credit risk assessments and fraud detection. The data science team needs a tool to monitor and track the performance and usage of these models hosted on endpoints, ensuring that they can easily manage and troubleshoot any issues related to real-time inference. To streamline this process, they are exploring which Amazon SageMaker service can help them track and manage their deployed models effectively. What do you recommend?