Azure Machine Learning now offers logging for web services RRS feed

  • General discussion

  • With the December 18th release of Azure Machine Learning, we have enabled logging to provide better diagnostics information for Web Service failures. Lack of clear error messages has been a key friction point for usage of operationalized services based on feedback we have seen in the forum and in the support aliases:

      • Logging can be enabled from the Web Service Dashboard page by clicking "Yes" then "Save". The “Learn more” linked page will be updated with help content shortly.
      • When Logging is enabled, all the diagnostics/errors from the default endpoint are logged to the Azure Storage Account linked with your workspace. You can see this storage account in the Azure portal Dashboard view (bottom of the Quick Glance section) of the workspace.
      • We enable Logging for the Web Service’s default endpoint only. We will soon be offering this feature for the additional endpoints that the user can create in the Azure portal.
      • The logs can be viewed using any of the several tools available to ‘explore’ a Storage Account. The easiest may be to simply navigate to the Storage Account in the Azure portal and then click on the CONTAINERS tab. You would then see a Container named ml-diagnostics. This container holds all the diagnostics info for all the web service endpoints for all the workspaces associated with this Storage account.
      • Each blob in the container holds the diagnostics info for exactly one of the following:
        1. An execution of the Batch-Execution method
        2. An execution of the Request-Response method
        3. Initialization of a Request-Response container
      • The name of each blob has a prefix of the following form: {Workspace Id}-{Web service Id}-{Endpoint Id}/{Log type}
      • Log Type takes one of the following values: batch, score/requests, or score/init
    Friday, December 19, 2014 9:19 PM