Azure Machine Learning data collector provides real-time logging of input and output data from models that are deployed to managed online endpoints or Kubernetes online endpoints. This data can be seamlessly used for model monitoring, debugging, or auditing, thereby providing observability into the performance of your deployed models.