Monitoring Azure Machine Learning deployments

You have successfully registered a trained model, an environment, a scoring file, and an inference configuration in the previous section. You have optimized your model for scoring and deployed it to a managed Kubernetes cluster. You auto-generated client SDKs for your ML services. So, can you finally lean back and enjoy the success of your hard work? Well, not yet! First, we need to make sure that we have all our monitoring in place so that you can observe and react to anything happening to your deployment.

First, the good things: with Azure Machine Learning deployments and managed compute targets, you will get many things included out of the box with either Azure, Azure Machine Learning, or your service used as a compute target. Tools such as the Azure dashboard, Azure Monitor, and Azure Log Analytics make it really easy to centralize log and debug information. Once your data is available through Log Analytics, it can be queried, analyzed...