It’s great because now it’s kind of unintuitive and complicated and it should be simple… I don’t understand it at all. I’m dealing with AI and just want to monitor my “endpoint”.
I used the workaround as mentioned by @ip97tpetyzd55d29 (thank you) for adding an “endpoint” and it seemed to work OK.
However I’ve really struggled to try and add metrics for “deployments” that hangs off the “endpoint”. It should be in the metric namespace Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments
Has anyone had any joy trying this? Thanks
How far did you get with your own monitoring @maciejglowacki or did you give up on it?
(I’m using Grafana Cloud)
Classic for me - as soon as I ask a question I discover the answer from re-reading everything again piece by piece
To help I’ll summarise below
To monitor an Endpoint…
Subscription: yourAzureSubscriptionID
Namespace: microsoft.machinelearningservices/workspaces/onlineendpoints
Region: yourAzureResourceRegion
Resource Group :yourAzureResourceGroup
Resource Name: yourAMLWorkspaceName/yourEndpointName
To monitor a Deployment (connected to a endpoint)…
Subscription: yourAzureSubscriptionID
Namespace: microsoft.machinelearningservices/workspaces/onlineendpoints/deployments
Region: yourAzureResourceRegion
Resource Group : yourAzureResourceGroup
Resource Name: yourAMLWorkspaceName/yourEndpointName/yourDeploymentName
1 Like