Supported metrics for Microsoft.CognitiveServices/accounts

The following table lists the metrics available for the Microsoft.CognitiveServices/accounts resource type.

Table headings

Metric - The metric display name as it appears in the Azure portal.
Name in Rest API - Metric name as referred to in the REST API.
Unit - Unit of measure.
Aggregation - The default aggregation type. Valid values: Average, Minimum, Maximum, Total, Count.
Dimensions - Dimensions available for the metric.
Time Grains - Intervals at which the metric is sampled. For example, PT1M indicates that the metric is sampled every minute, PT30M every 30 minutes, PT1H every hour, and so on.
DS Export- Whether the metric is exportable to Azure Monitor Logs via Diagnostic Settings.

For information on exporting metrics, see - Metrics export using data collection rules and Create diagnostic settings in Azure Monitor.

For information on metric retention, see Azure Monitor Metrics overview.

For a list of supported logs, see Supported log categories - Microsoft.CognitiveServices/accounts

Category Metric Name in REST API Unit Aggregation Dimensions Time Grains DS Export
Feature Occurences Action Feature Occurrences

Number of times each action feature appears.
ActionFeatureIdOccurrences Count Total (Sum) FeatureId, Mode, RunId PT1M Yes
Features Per Event Action Features Per Event

Average number of action features per event.
ActionFeaturesPerEvent Count Average Mode, RunId PT1M Yes
Actions Action Occurences

Number of times each action appears.
ActionIdOccurrences Count Total (Sum) ActionId, Mode, RunId PT1M Yes
Namespaces Per Event Action Namespaces Per Event

Average number of action namespaces per event.
ActionNamespacesPerEvent Count Average Mode, RunId PT1M Yes
Actions Actions Per Event

Number of actions per event.
ActionsPerEvent Count Average Mode, RunId PT1M Yes
Azure OpenAI - Usage Active Tokens

Total tokens minus cached tokens over a period of time. Applies to PTU and PTU-managed deployments. Use this metric to understand your TPS or TPM based utilization for PTUs and compare to your benchmarks for target TPS or TPM for your scenarios. To breakdown API requests, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName, and ModelVersion.
ActiveTokens Count Minimum, Maximum, Average, Total (Sum) Region, ModelDeploymentName, ModelName, ModelVersion PT1M Yes
SpeechServices - Usage Audio Seconds Transcribed

Number of seconds transcribed
AudioSecondsTranscribed Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
SpeechServices - Usage Audio Seconds Translated

Number of seconds translated
AudioSecondsTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
SpeechServices - Usage Avatar Model Hosting Seconds

Number of Seconds.
AvatarModelHostingSeconds Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Azure OpenAI - HTTP Requests Prompt Token Cache Match Rate

Percentage of the prompt tokens hit the cache, avaiable for PTU-managed.
AzureOpenAIContextTokensCacheMatchRate Percent Minimum, Maximum, Average Region, ModelDeploymentName, ModelName, ModelVersion PT1M No
Azure OpenAI - Usage Provisioned-managed Utilization

Utilization % for a provisoned-managed deployment, calculated as (PTUs consumed / PTUs deployed) x 100. When utilization is greater than or equal to 100%, calls are throttled and error code 429 returned. To breakdown this metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName, ModelVersion and StreamType (Streaming vs non-streaming requests)
AzureOpenAIProvisionedManagedUtilization Percent Minimum, Maximum, Average Region, StreamType, ModelDeploymentName, ModelName, ModelVersion PT1M No
Azure OpenAI - Usage Provisioned-managed Utilization V2

Utilization % for a provisoned-managed deployment, calculated as (PTUs consumed / PTUs deployed) x 100. When utilization is greater than or equal to 100%, calls are throttled and error code 429 returned. To breakdown this metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName, ModelVersion and StreamType (Streaming vs non-streaming requests)
AzureOpenAIProvisionedManagedUtilizationV2 Percent Minimum, Maximum, Average Region, StreamType, ModelDeploymentName, ModelName, ModelVersion PT1M No
Azure OpenAI - HTTP Requests Azure OpenAI Requests

Number of calls made to the Azure OpenAI API over a period of time. Applies to PTU, PTU-Managed and Pay-as-you-go deployments. To breakdown API requests, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName, ModelVersion, StatusCode (successful, clienterrors, server errors), StreamType (Streaming vs non-streaming requests) and operation.
AzureOpenAIRequests Count Total (Sum) ApiName, OperationName, Region, StreamType, ModelDeploymentName, ModelName, ModelVersion, StatusCode PT1M Yes
Azure OpenAI - HTTP Requests Time to Response

Recommended latency (responsiveness) measure for streaming requests. Applies to PTU and PTU-managed deployments. Calculated as time taken for the first response to appear after a user sends a prompt, as measured by the API gateway. This number increases as the prompt size increases and/or cache hit size reduces. To breakdown time to response metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName, and ModelVersion.

Note: this metric is an approximation as measured latency is heavily dependent on multiple factors, including concurrent calls and overall workload pattern. In addition, it does not account for any client-side latency that may exist between your client and the API endpoint. Please refer to your own logging for optimal latency tracking.
AzureOpenAITimeToResponse MilliSeconds Minimum, Maximum, Average ApiName, OperationName, Region, StreamType, ModelDeploymentName, ModelName, ModelVersion, StatusCode PT1M Yes
Slots Baseline Estimator Overall Reward

Baseline Estimator Overall Reward.
BaselineEstimatorOverallReward Count Average Mode, RunId PT1M Yes
Slots Baseline Estimator Slot Reward

Baseline Estimator Reward by slot.
BaselineEstimatorSlotReward Count Average SlotId, SlotIndex, Mode, RunId PT1M Yes
Slots Baseline Random Estimator Overall Reward

Baseline Random Estimator Overall Reward.
BaselineRandomEstimatorOverallReward Count Average Mode, RunId PT1M Yes
Slots Baseline Random Estimator Slot Reward

Baseline Random Estimator Reward by slot.
BaselineRandomEstimatorSlotReward Count Average SlotId, SlotIndex, Mode, RunId PT1M Yes
Estimations Baseline Random Event count

Estimation for baseline random event count.
BaselineRandomEventCount Count Total (Sum) Mode, RunId PT1M Yes
Estimations Baseline Random Reward

Estimation for baseline random reward.
BaselineRandomReward Count Total (Sum) Mode, RunId PT1M Yes
Cognitive Services - HTTP Requests Blocked Calls

Number of calls that exceeded rate or quota limit. Do not use for Azure OpenAI service.
BlockedCalls Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Usage Inference Count

Inference Count of Carnegie Frontdoor Service
CarnegieInferenceCount Count Total (Sum) Region, Modality, Category, Language, SeverityLevel, UseCustomList PT1M Yes
Usage Characters Trained (Deprecated)

Total number of characters trained.
CharactersTrained Count Total (Sum) ApiName, OperationName, Region PT1M Yes
Usage Characters Translated (Deprecated)

Total number of characters in incoming text request.
CharactersTranslated Count Total (Sum) ApiName, OperationName, Region PT1M Yes
Cognitive Services - HTTP Requests Client Errors

Number of calls with client side error (HTTP response code 4xx). Do not use for Azure OpenAI service.
ClientErrors Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Usage Computer Vision Transactions

Number of Computer Vision Transactions
ComputerVisionTransactions Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
ContentSafety - Usage Call Count for Image Moderation

Number of calls for image moderation.
ContentSafetyImageAnalyzeRequestCount Count Total (Sum) ApiVersion PT1M Yes
ContentSafety - Usage Call Count for Text Moderation

Number of calls for text moderation.
ContentSafetyTextAnalyzeRequestCount Count Total (Sum) ApiVersion PT1M Yes
Feature Occurences Context Feature Occurrences

Number of times each context feature appears.
ContextFeatureIdOccurrences Count Total (Sum) FeatureId, Mode, RunId PT1M Yes
Features Per Event Context Features Per Event

Number of context features per event.
ContextFeaturesPerEvent Count Average Mode, RunId PT1M Yes
Namespaces Per Event Context Namespaces Per Event

Number of context namespaces per event.
ContextNamespacesPerEvent Count Average Mode, RunId PT1M Yes
Usage Custom Vision Training Time

Custom Vision training time
CustomVisionTrainingTime Seconds Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Custom Vision Transactions

Number of Custom Vision prediction transactions
CustomVisionTransactions Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Cognitive Services - HTTP Requests Data In

Size of incoming data in bytes. Do not use for Azure OpenAI service.
DataIn Bytes Total (Sum) ApiName, OperationName, Region PT1M Yes
Cognitive Services - HTTP Requests Data Out

Size of outgoing data in bytes. Do not use for Azure OpenAI service.
DataOut Bytes Total (Sum) ApiName, OperationName, Region PT1M Yes
Usage Document Characters Translated

Number of characters in document translation request.
DocumentCharactersTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Document Custom Characters Translated

Number of characters in custom document translation request.
DocumentCustomCharactersTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Face Images Trained

Number of images trained. 1,000 images trained per transaction.
FaceImagesTrained Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Faces Stored

Number of faces stored, prorated daily. The number of faces stored is reported daily.
FacesStored Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Face Transactions

Number of API calls made to Face service
FaceTransactions Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
FeatureCardinality Feature Cardinality by Action

Feature Cardinality based on Action.
FeatureCardinality_Action Count Average FeatureId, Mode, RunId PT1M Yes
FeatureCardinality Feature Cardinality by Context

Feature Cardinality based on Context.
FeatureCardinality_Context Count Average FeatureId, Mode, RunId PT1M Yes
FeatureCardinality Feature Cardinality by Slot

Feature Cardinality based on Slot.
FeatureCardinality_Slot Count Average FeatureId, Mode, RunId PT1M Yes
Azure OpenAI - Usage Processed FineTuned Training Hours

Number of Training Hours Processed on an OpenAI FineTuned Model
FineTunedTrainingHours Count Total (Sum) ApiName, ModelDeploymentName, FeatureName, UsageChannel, Region PT1M Yes
Azure OpenAI - Usage Generated Completion Tokens

Number of tokens generated (output) from an OpenAI model. Applies to PTU, PTU-Managed and Pay-as-you-go deployments. To breakdown this metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName and ModelName.
GeneratedTokens Count Total (Sum) ApiName, ModelDeploymentName, FeatureName, UsageChannel, Region, ModelVersion PT1M Yes
Usage Images Stored

Number of Custom Vision images stored.
ImagesStored Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Cognitive Services - HTTP Requests Latency

Latency in milliseconds. Do not use for Azure OpenAI service.
Latency MilliSeconds Average ApiName, OperationName, Region, RatelimitKey PT1M Yes
Usage Learned Events

Number of Learned Events.
LearnedEvents Count Total (Sum) IsMatchBaseline, Mode, RunId PT1M Yes
Usage LUIS Speech Requests

Number of LUIS speech to intent understanding requests
LUISSpeechRequests Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage LUIS Text Requests

Number of LUIS text requests
LUISTextRequests Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Matched Rewards

Number of Matched Rewards.
MatchedRewards Count Total (Sum) Mode, RunId PT1M Yes
Usage Non Activated Events

Number of skipped events.
NonActivatedEvents Count Total (Sum) Mode, RunId PT1M Yes
Slots Slots

Number of slots per event.
NumberOfSlots Count Average Mode, RunId PT1M Yes
SpeechServices - Usage Number of Speaker Profiles

Number of speaker profiles enrolled. Prorated hourly.
NumberofSpeakerProfiles Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Observed Rewards

Number of Observed Rewards.
ObservedRewards Count Total (Sum) Mode, RunId PT1M Yes
Usage Document Sync Characters Translated

Number of characters in document translation (synchronous) request.
OneDocumentCharactersTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Slots Online Estimator Overall Reward

Online Estimator Overall Reward.
OnlineEstimatorOverallReward Count Average Mode, RunId PT1M Yes
Slots Online Estimator Slot Reward

Online Estimator Reward by slot.
OnlineEstimatorSlotReward Count Average SlotId, SlotIndex, Mode, RunId PT1M Yes
Estimations Online Event Count

Estimation for online event count.
OnlineEventCount Count Total (Sum) Mode, RunId PT1M Yes
Estimations Online Reward

Estimation for online reward.
OnlineReward Count Total (Sum) Mode, RunId PT1M Yes
Usage Processed Characters

Number of Characters processed by Immersive Reader.
ProcessedCharacters Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Processed Health Text Records

Number of health text records processed
ProcessedHealthTextRecords Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Processed Images

Number of images processed
ProcessedImages Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Processed Pages

Number of pages processed
ProcessedPages Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Azure OpenAI - Usage Processed Prompt Tokens

Number of prompt tokens processed (input) on an OpenAI model. Applies to PTU, PTU-Managed and Pay-as-you-go deployments. To breakdown this metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName and ModelName.
ProcessedPromptTokens Count Total (Sum) ApiName, ModelDeploymentName, FeatureName, UsageChannel, Region, ModelVersion PT1M Yes
Usage Processed Text Records

Count of Text Records.
ProcessedTextRecords Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage QA Text Records

Number of text records processed
QuestionAnsweringTextRecords Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
ContentSafety - Risks&Safety Potentially Abusive User Count

Number of potentially abusive user that detected over a period of time. You can add a filter or apply splitting by the following dimension: ModelDeploymentName.
RAIAbusiveUsersCount Count Total (Sum) Region, ModelDeploymentName PT1M Yes
ContentSafety - Risks&Safety Harmful Volume Detected

Number of calls made to Azure OpenAI API and detected as harmful(both block model and annotate mode) by content filter applied over a period of time. You can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName and TextType.
RAIHarmfulRequests Count Total (Sum) Region, ModelDeploymentName, ModelName, ModelVersion, ApiName, TextType, Category, Severity PT1M Yes
ContentSafety - Risks&Safety Blocked Volume

Number of calls made to Azure OpenAI API and rejected by content filter applied over a period of time. You can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName and TextType.
RAIRejectedRequests Count Total (Sum) Region, ModelDeploymentName, ModelName, ModelVersion, ApiName, TextType, Category PT1M Yes
ContentSafety - Risks&Safety Safety System Event

System event for risks & safety monitoring. You can add a filter or apply splitting by the following dimension: EventType.
RAISystemEvent Count Average Region, EventType PT1M Yes
ContentSafety - Risks&Safety Total Volume Sent For Safety Check

Number of calls made to Azure OpenAI API and detected by content filter applied over a period of time. You can add a filter or apply splitting by the following dimensions: ModelDeploymentName, ModelName.
RAITotalRequests Count Total (Sum) Region, ModelDeploymentName, ModelName, ModelVersion, ApiName PT1M Yes
Cognitive Services - HTTP Requests Ratelimit

The current ratelimit of the ratelimit key. Do not use for Azure OpenAI service.
Ratelimit Count Total (Sum) Region, RatelimitKey PT1M Yes
Rewards Average Reward Per Event

Average reward per event.
Reward Count Average BaselineAction, ChosenActionId, MatchesBaseline, NonDefaultReward, Mode, RunId PT1M Yes
Cognitive Services - HTTP Requests Server Errors

Number of calls with service internal error (HTTP response code 5xx). Do not use for Azure OpenAI service.
ServerErrors Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Feature Occurences Slot Feature Occurrences

Number of times each slot feature appears.
SlotFeatureIdOccurrences Count Total (Sum) FeatureId, Mode, RunId PT1M Yes
Features Per Event Slot Features Per Event

Average number of slot features per event.
SlotFeaturesPerEvent Count Average Mode, RunId PT1M Yes
Slots Slot Occurrences

Number of times each slot appears.
SlotIdOccurrences Count Total (Sum) SlotId, SlotIndex, Mode, RunId PT1M Yes
Namespaces Per Event Slot Namespaces Per Event

Average number of slot namespaces per event.
SlotNamespacesPerEvent Count Average Mode, RunId PT1M Yes
Rewards Slot Reward

Reward per slot.
SlotReward Count Average BaselineActionId, ChosenActionId, MatchesBaseline, NonDefaultReward, SlotId, SlotIndex, Mode, RunId PT1M Yes
SpeechServices - Usage Speaker Recognition Transactions

Number of speaker recognition transactions
SpeakerRecognitionTransactions Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
SpeechServices - Usage Speech Model Hosting Hours

Number of speech model hosting hours
SpeechModelHostingHours Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Speech Session Duration (Deprecated)

Total duration of speech session in seconds.
SpeechSessionDuration Seconds Total (Sum) ApiName, OperationName, Region PT1M Yes
Cognitive Services - HTTP Requests Successful Calls

Number of successful calls. Do not use for Azure OpenAI service.
SuccessfulCalls Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Cognitive Services - SLI AvailabilityRate

Availability percentage with the following calculation: (Total Calls - Server Errors)/Total Calls. Server Errors include any HTTP responses >=500. Do not use for Azure OpenAI service.
SuccessRate Percent Minimum, Maximum, Average ApiName, OperationName, Region, RatelimitKey PT1M No
SpeechServices - Usage Synthesized Characters

Number of Characters.
SynthesizedCharacters Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Text Characters Translated

Number of characters in incoming text translation request.
TextCharactersTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Text Custom Characters Translated

Number of characters in incoming custom text translation request.
TextCustomCharactersTranslated Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Usage Text Trained Characters

Number of characters trained using text translation.
TextTrainedCharacters Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
Azure OpenAI - Usage Processed Inference Tokens

Number of inference tokens processed on an OpenAI model. Calculated as prompt tokens (input) plus generated tokens (output). Applies to PTU, PTU-Managed and Pay-as-you-go deployments. To breakdown this metric, you can add a filter or apply splitting by the following dimensions: ModelDeploymentName and ModelName.
TokenTransaction Count Total (Sum) ApiName, ModelDeploymentName, FeatureName, UsageChannel, Region, ModelVersion PT1M Yes
Cognitive Services - HTTP Requests Total Calls

Total number of calls. Do not use for Azure OpenAI service.
TotalCalls Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Cognitive Services - HTTP Requests Total Errors

Total number of calls with error response (HTTP response code 4xx or 5xx). Do not use for Azure OpenAI service.
TotalErrors Count Total (Sum) ApiName, OperationName, Region, RatelimitKey PT1M Yes
Usage Total Events

Number of events.
TotalEvents Count Total (Sum) Mode, RunId PT1M Yes
Cognitive Services - HTTP Requests Total Token Calls

Total number of token calls.
TotalTokenCalls Count Total (Sum) ApiName, OperationName, Region PT1M Yes
Usage Total Transactions (Deprecated)

Total number of transactions.
TotalTransactions Count Total (Sum) <none> PT1M Yes
Estimations User Baseline Event Count

Estimation for user defined baseline event count.
UserBaselineEventCount Count Total (Sum) Mode, RunId PT1M Yes
Estimations User Baseline Reward

Estimation for user defined baseline reward.
UserBaselineReward Count Total (Sum) Mode, RunId PT1M Yes
SpeechServices - Usage Video Seconds Synthesized

Number of seconds synthesized
VideoSecondsSynthesized Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
SpeechServices - Usage Voice Model Hosting Hours

Number of Hours.
VoiceModelHostingHours Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes
SpeechServices - Usage Voice Model Training Minutes

Number of Minutes.
VoiceModelTrainingMinutes Count Total (Sum) ApiName, FeatureName, UsageChannel, Region PT1M Yes

Next steps