Monitoring AzureML Batch Requests

Question

I'm currently trying to find out if there is a possibility to monitor the following things for an AzureML Webservice

- which batch requests (ID) are currently running

- how long has a batch job been running

Does anyone know if there is something like a REST API for this?

Or is there any possibility at all to accomplish what I'm looking for?

Answer 1

Hi jbreunung,

Would enable logging for the web service help you to monitor your Azure ML web services?

https://social.msdn.microsoft.com/Forums/azure/en-US/6fc5f8cf-0ab7-4177-881b-19408826554d/azure-machine-learning-now-offers-logging-for-web-services?forum=MachineLearning

Regards,
Jaya.

Answer 2

You can try this REST API call to get job details:

Request:

GET

https://ussouthcentral.services.azureml.net/workspaces/<id>/services/<id>/jobs?api-version=2.0

Authorization: Bearer <key>

Response:

[{"Id":"id","WorkspaceId":"id","WebServiceId":"id","CreatedAt":"2015-12-16T17:09:30.5166869Z","StartTime":"2015-12-16T17:09:34.231Z","EndTime":"2015-12-16T17:09:40.649Z","StatusCode":"Finished"}

]

Thanks,

Raymond

Answer 3

That is what I was looking for. Thank you!

Is there any kind of documentation about this api call?

e.g. how many batch jobs are returned in the response, are older jobs automatically removed from the list etc.?