Azure Machine Learning monitoring

  • How-to guide
  • 5-min read
  • Published Aug 19, 2020

Dynatrace ingests metrics for multiple preselected namespaces, including Azure Machine Learning. You can view metrics for each service instance, split metrics into multiple dimensions, and create custom charts that you can pin to your dashboards.

Prerequisites

  • Dynatrace version 1.200+
  • Environment ActiveGate version 1.195+

Enable monitoring

To learn how to enable service monitoring, see Enable service monitoring.

View service metrics

You can view the service metrics in your Dynatrace environment either on the custom device overview page or on your Dashboards page.

View metrics on the custom device overview page

To access the custom device overview page

  1. Go to Technologies & Processes Classic.
  2. Filter by service name and select the relevant custom device group.
  3. Once you select the custom device group, you're on the custom device group overview page.
  4. The custom device group overview page lists all instances (custom devices) belonging to the group. Select an instance to view the custom device overview page.

View metrics on your dashboard

Once you add a service to monitoring, a preset dashboard for the respective service containing all recommended metrics is automatically created on your Dashboards page. You can look for specific dashboards by filtering by Preset and then by Name.

For existing monitored services, you might need to resave your credentials for the preset dashboard to appear on the Dashboards page. To resave your credentials, go to Settings > Cloud and virtualization > Azure, select the desired Azure instance, then select Save.

You can't make changes on a preset dashboard directly, but you can clone and edit it. To clone a dashboard, open the browse menu () and select Clone.
To remove a dashboard from the dashboards list, you can hide it. To hide a dashboard, open the browse menu () and select Hide.

Hiding a dashboard doesn't affect other users.

Clone hide azure

Machine

Learning

Available metrics

NameDescriptionDimensionsUnitRecommended
Active CoresNumber of active cores.Scenario, ClusterNameCountApplicable
Active NodesNumber of active nodes. These are the nodes which are actively running a job.Scenario, ClusterNameCountApplicable
Cancel Requested RunsNumber of runs where cancel was requested for this workspace. Count is updated when cancellation request has been received for a run.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCount
Cancelled RunsNumber of runs cancelled for this workspace. Count is updated when a run is successfully cancelled.Scenario, RunType, PublishedPipelineI, ComputeType, PipelineStepTypeCount
Completed RunsNumber of runs completed successfully for this workspace. Count is updated when a run has completed and output has been collected.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
CpuUtilizationPercentage of memory utilization on a CPU node. Utilization is reported at one minute intervals.Scenario, runId, NodeId, ClusterNamePercentApplicable
ErrorsNumber of run errors in this workspace. Count is updated whenever run encounters an error.ScenarioCountApplicable
Failed RunsNumber of runs failed for this workspace. Count is updated when a run fails.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Finalizing RunsNumber of runs entered finalizing state for this workspace. Count is updated when a run has completed but output collection still in progress.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
GpuUtilizationPercentage of memory utilization on a GPU node. Utilization is reported at one-minute intervals.Scenario, runId, NodeId, DeviceId, ClusterNamePercentApplicable
Idle CoresNumber of idle cores.Scenario, ClusterNameCountApplicable
Idle NodesNumber of idle nodes. Idle nodes are the nodes which are not running any jobs but can accept new job if available.Scenario, ClusterNameCountApplicable
Leaving CoresNumber of leaving coresScenario, ClusterNameCountApplicable
Leaving NodesNumber of leaving nodes. Leaving nodes are the nodes which just finished processing a job and will go to Idle state.Scenario, ClusterNameCountApplicable
Model Deploy FailedNumber of model deployments that failed in this workspace.Scenario, StatusCodeCountApplicable
Model Deploy StartedNumber of model deployments started in this workspace.ScenarioCountApplicable
Model Deploy SucceededNumber of model deployments that succeeded in this workspace.ScenarioCountApplicable
Model Register FailedNumber of model registrations that failed in this workspace.Scenario, StatusCodeCountApplicable
Model Register SucceededNumber of model registrations that succeeded in this workspace.ScenarioCountApplicable
Not Responding RunsNumber of runs not responding for this workspace. Count is updated when a run enters Not Responding state.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Not Started RunsNumber of runs in Not Started state for this workspace. Count is updated when a request is received to create a run but run information has not yet been populated.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Preempted CoresNumber of preempted coresScenario, ClusterNameCountApplicable
Preempted NodesNumber of preempted nodes. These nodes are the low priority nodes which are taken away from the available node pool.Scenario, ClusterNameCountApplicable
Preparing RunsNumber of runs that are preparing for this workspace. Count is updated when a run enters Preparing state while the run environment is being prepared.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCount
Provisioning RunsNumber of runs that are provisioning for this workspace. Count is updated when a run is waiting on compute target creation or provisioning.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCount
Queued RunsNumber of runs that are queued for this workspace. Count is updated when a run is queued in compute target. Can occur when waiting for required compute nodes to be ready.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Quota Utilization PercentagePercent of quota utilized.Scenario, ClusterName, VmFamilyName, VmPriorityPercentApplicable
Started RunsNumber of runs running for this workspace. Count is updated when a run starts running on required resources.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Starting RunsNumber of runs started for this workspace. Count is updated after request to create run and run info, such as the Run Id, has been populated.Scenario, RunType, PublishedPipelineId, ComputeType, PipelineStepTypeCountApplicable
Total CoresNumber of total cores.Scenario, ClusterNameCountApplicable
Total NodesNumber of total nodes. This total includes some of Active Nodes, Idle Nodes, Unusable Nodes, Preempted Nodes, Leaving Nodes.Scenario, ClusterNameCountApplicable
Unusable CoresNumber of unusable cores.Scenario, ClusterNameCountApplicable
Unusable NodesNumber of unusable nodes. Unusable nodes are not functional due to some unresolvable issue. Azure will recycle these nodes.Scenario, ClusterNameCountApplicable
WarningsNumber of run warnings in this workspace. Count is updated whenever a run encounters a warning.ScenarioCountApplicable
Related tags
Infrastructure Observability