Open pietz opened 2 years ago
Apparently, the workflow has changed with real-time and batch online endpoints. Microsoft has recently updated their docs accordingly: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-monitor-online-endpoints
Other than that I found it's also possible to use the AzureLogHandler
and a separate instance of Application Insights to log custom messages.
Which example? Describe the issue
example: online-endpoints-simple-deployment.ipynb
description: My code is based on this notebook. However, I added the
app_insights_enabled
flag to the deployment variable in order to save logs generated in score.py to application insights. This doesn't work. I tried print statements and logging.info statements but I cannot see any of these entries in application insights of the deployment.How can I solve this problem? Or what's the recommended way to save custom logging commands to an azure logging service?