--- title: Azure Cognitive Services description: Track key Azure Cognitive Services metrics. breadcrumbs: Docs > Integrations > Azure Cognitive Services --- > For the complete documentation index, see [llms.txt](https://docs.datadoghq.com/llms.txt). # Azure Cognitive Services {% callout %} # Important note for users on the following Datadog sites: us2.ddog-gov.com {% alert level="info" %} To find out if this integration is available in your organization, see your [Datadog Integrations](https://app.datadoghq.com/integrations) page or ask your organization administrator. To initiate an exception request to enable this integration for your organization, email [support@ddog-gov.com](mailto:support@ddog-gov.com). {% /alert %} {% /callout %} ## Overview{% #overview %} [Azure Cognitive Services](https://learn.microsoft.com/azure/ai-services/what-are-ai-services) are APIs, SDKs, and services that help developers build intelligent applications without direct AI or data science skills. Use the Datadog Azure integration to collect metrics from Azure Cognitive Services. ## Setup{% #setup %} ### Installation{% #installation %} If you haven't already, set up the [Microsoft Azure integration](https://docs.datadoghq.com/integrations/azure.md). No additional installation are required. ## Data Collected{% #data-collected %} ### Metrics{% #metrics %} | | | | | **azure.cognitiveservices_accounts.action_feature_occurrences**(count) | Number of times each action feature appears. | | **azure.cognitiveservices_accounts.action_features_per_event**(gauge) | Average number of action features per event. | | **azure.cognitiveservices_accounts.action_occurences**(count) | Number of times each action appears. | | **azure.cognitiveservices_accounts.action_namespaces_per_event**(gauge) | Average number of action namespaces per event. | | **azure.cognitiveservices_accounts.actions_per_event**(gauge) | Number of actions per event. | | **azure.cognitiveservices_accounts.audio_seconds_transcribed**(count) | Number of seconds transcribed. | | **azure.cognitiveservices_accounts.audio_seconds_translated**(count) | Number of seconds translated. | | **azure.cognitiveservices_accounts.baseline_estimator_overall_reward**(gauge) | Baseline estimator overall reward. | | **azure.cognitiveservices_accounts.baseline_estimator_slot_reward**(gauge) | Baseline estimator reward by slot. | | **azure.cognitiveservices_accounts.baseline_random_estimator_overall_reward**(gauge) | Baseline random estimator overall reward. | | **azure.cognitiveservices_accounts.baseline_random_estimator_slot_reward**(gauge) | Baseline random estimator reward by slot. | | **azure.cognitiveservices_accounts.baseline_random_event_count**(count) | Estimation for baseline random event count. | | **azure.cognitiveservices_accounts.baseline_random_reward**(count) | Estimation for baseline random reward. | | **azure.cognitiveservices_accounts.blocked_calls**(count) | Number of calls that exceeded rate or quota limit. | | **azure.cognitiveservices_accounts.inference_count**(count) | Inference count of carnegie frontdoor service. | | **azure.cognitiveservices_accounts.characters_trained_deprecated**(count) | Total number of characters trained. | | **azure.cognitiveservices_accounts.characters_translated_deprecated**(count) | Total number of characters in incoming text request. | | **azure.cognitiveservices_accounts.client_errors**(count) | Number of calls with client side error (HTTP response code 4xx). | | **azure.cognitiveservices_accounts.computer_vision_transactions**(count) | Number of computer vision transactions. | | **azure.cognitiveservices_accounts.context_feature_occurrences**(count) | Number of times each context feature appears. | | **azure.cognitiveservices_accounts.context_features_per_event**(gauge) | Number of context features per event. | | **azure.cognitiveservices_accounts.context_namespaces_per_event**(gauge) | Number of context namespaces per event. | | **azure.cognitiveservices_accounts.custom_vision_training_time**(count) | Custom vision training time.*Shown as second* | | **azure.cognitiveservices_accounts.custom_vision_transactions**(count) | Number of custom vision prediction transactions. | | **azure.cognitiveservices_accounts.data_in**(count) | Size of incoming data in bytes.*Shown as byte* | | **azure.cognitiveservices_accounts.data_out**(count) | Size of outgoing data in bytes.*Shown as byte* | | **azure.cognitiveservices_accounts.document_characters_translated**(count) | Number of characters in document translation request. | | **azure.cognitiveservices_accounts.document_custom_characters_translated**(count) | Number of characters in custom document translation request. | | **azure.cognitiveservices_accounts.face_images_trained**(count) | Number of images trained. 1,000 images trained per transaction. | | **azure.cognitiveservices_accounts.faces_stored**(count) | Number of faces stored, prorated daily. The number of faces stored is reported daily. | | **azure.cognitiveservices_accounts.face_transactions**(count) | Number of API calls made to face service. | | **azure.cognitiveservices_accounts.feature_cardinality_by_action**(gauge) | Feature cardinality based on action. | | **azure.cognitiveservices_accounts.feature_cardinality_by_context**(gauge) | Feature cardinality based on context. | | **azure.cognitiveservices_accounts.feature_cardinality_by_slot**(gauge) | Feature cardinality based on slot. | | **azure.cognitiveservices_accounts.generated_completion_tokens**(count) | Number of Generated Completion Tokens from an OpenAI model. | | **azure.cognitiveservices_accounts.processed_fine_tuned_training_hours**(count) | Number of training hours processed on an OpenAI finetuned model. | | **azure.cognitiveservices_accounts.processed_prompt_tokens**(count) | Number of Prompt Tokens Processed on an OpenAI model | | **azure.cognitiveservices_accounts.images_stored**(count) | Number of custom vision images stored. | | **azure.cognitiveservices_accounts.latency**(gauge) | Latency in milliseconds.*Shown as millisecond* | | **azure.cognitiveservices_accounts.learned_events**(count) | Number of learned events. | | **azure.cognitiveservices_accounts.luis_speech_requests**(count) | Number of LUIS speech to intent understanding requests. | | **azure.cognitiveservices_accounts.luis_text_requests**(count) | Number of LUIS text requests. | | **azure.cognitiveservices_accounts.matched_rewards**(count) | Number of matched rewards. | | **azure.cognitiveservices_accounts.non_activated_events**(count) | Number of skipped events. | | **azure.cognitiveservices_accounts.slots**(gauge) | Number of slots per event. | | **azure.cognitiveservices_accounts.number_of_speaker_profiles**(count) | Number of speaker profiles enrolled. Prorated hourly. | | **azure.cognitiveservices_accounts.observed_rewards**(count) | Number of observed rewards. | | **azure.cognitiveservices_accounts.online_estimator_overall_reward**(gauge) | Online estimator overall reward. | | **azure.cognitiveservices_accounts.online_estimator_slot_reward**(gauge) | Online estimator reward by slot. | | **azure.cognitiveservices_accounts.online_event_count**(count) | Estimation for online event count. | | **azure.cognitiveservices_accounts.online_reward**(count) | Estimation for online reward. | | **azure.cognitiveservices_accounts.processed_characters**(count) | Number of characters processed by immersive reader. | | **azure.cognitiveservices_accounts.processed_health_text_records**(count) | Number of health text records processed. | | **azure.cognitiveservices_accounts.processed_images**(count) | Number of images processed. | | **azure.cognitiveservices_accounts.processed_pages**(count) | Number of pages processed. | | **azure.cognitiveservices_accounts.processed_text_records**(count) | Count of text records. | | **azure.cognitiveservices_accounts.qa_text_records**(count) | Number of text records processed. | | **azure.cognitiveservices_accounts.ratelimit**(count) | Current rate limit of the ratelimit key. | | **azure.cognitiveservices_accounts.average_reward_per_event**(gauge) | Average reward per event. | | **azure.cognitiveservices_accounts.server_errors**(count) | Number of calls with service internal error (HTTP response code 5xx). | | **azure.cognitiveservices_accounts.slot_feature_occurrences**(count) | Number of times each slot feature appears. | | **azure.cognitiveservices_accounts.slot_features_per_event**(gauge) | Average number of slot features per event. | | **azure.cognitiveservices_accounts.slot_occurrences**(count) | Number of times each slot appears. | | **azure.cognitiveservices_accounts.slot_namespaces_per_event**(gauge) | Average number of slot namespaces per event. | | **azure.cognitiveservices_accounts.slot_reward**(gauge) | Reward per slot. | | **azure.cognitiveservices_accounts.speaker_recognition_transactions**(count) | Number of speaker recognition transactions. | | **azure.cognitiveservices_accounts.speech_model_hosting_hours**(count) | Number of speech model hosting hours. | | **azure.cognitiveservices_accounts.speech_session_duration_deprecated**(count) | Total duration of speech session in seconds.*Shown as second* | | **azure.cognitiveservices_accounts.successful_calls**(count) | Number of successful calls. | | **azure.cognitiveservices_accounts.availability**(gauge) | Availability percentage with the following calculation: (total calls - server errors)/total calls. Server errors include any HTTP responses greater than or equal to 500.*Shown as percent* | | **azure.cognitiveservices_accounts.synthesized_characters**(count) | Number of characters. | | **azure.cognitiveservices_accounts.text_characters_translated**(count) | Number of characters in incoming text translation request. | | **azure.cognitiveservices_accounts.text_custom_characters_translated**(count) | Number of characters in incoming custom text translation request. | | **azure.cognitiveservices_accounts.text_trained_characters**(count) | Number of characters trained using text translation. | | **azure.cognitiveservices_accounts.processed_inference_tokens**(count) | Number of inference tokens processed on an OpenAI model. | | **azure.cognitiveservices_accounts.total_calls**(count) | Total number of calls. | | **azure.cognitiveservices_accounts.total_errors**(count) | Total number of calls with error response (HTTP response code 4xx or 5xx). | | **azure.cognitiveservices_accounts.total_events**(count) | Number of events. | | **azure.cognitiveservices_accounts.total_token_calls**(count) | Total number of token calls. | | **azure.cognitiveservices_accounts.total_transactions_deprecated**(count) | Total number of transactions. | | **azure.cognitiveservices_accounts.user_baseline_event_count**(count) | Estimation for user-defined baseline event count. | | **azure.cognitiveservices_accounts.user_baseline_reward**(count) | Estimation for user-defined baseline reward. | | **azure.cognitiveservices_accounts.voice_model_hosting_hours**(count) | Number of hours. | | **azure.cognitiveservices_accounts.voice_model_training_minutes**(count) | Number of minutes. | | **azure.cognitiveservices_accounts.count**(gauge) | Count of CognitiveServices accounts. | | **azure.cognitiveservices_accounts.azure_open_ai_availability_rate**(gauge) | Availability percentage with the following calculation: (total calls - server errors)/total calls. Server errors include any HTTP responses >=500.*Shown as percent* | | **azure.cognitiveservices_accounts.time_between_token**(gauge) | For streaming requests; model token generation rate, measured in milliseconds. Applies to PTU, ptu-managed and pay-as-you-go deployments.*Shown as millisecond* | | **azure.cognitiveservices_accounts.normalized_time_to_first_byte**(gauge) | For streaming and non-streaming requests; time it takes for first byte of response data to be received after request is made by model, normalized by token. Applies to PTU, ptu-managed, and pay-as-you-go deployments.*Shown as millisecond* | | **azure.cognitiveservices_accounts.tokens_per_second**(gauge) | Generation speed of a model response, in tokens per second: total tokens generated divided by the time to generate them. Applies to PTU and PTU-managed deployments. | | **azure.cognitiveservices_accounts.time_to_last_byte**(gauge) | For streaming and non-streaming requests; time it takes for last byte of response data to be received after request is made by model. Applies to PTU, ptu-managed, and pay-as-you-go deployments.*Shown as millisecond* | | **azure.cognitiveservices_accounts.audio_completion_tokens**(count) | Number of audio prompt tokens generated (output) on an openai model. Applies to ptu-managed and pay-as-you-go model deployments. | | **azure.cognitiveservices_accounts.audio_prompt_tokens**(count) | Number of audio prompt tokens processed (input) on an openai model. Applies to ptu-managed and pay-as-you-go model deployments. | | **azure.cognitiveservices_accounts.provisioned_managed_utilization_deprecated**(gauge) | 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).*Shown as percent* | | **azure.cognitiveservices_accounts.availability_rate**(gauge) | 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.*Shown as percent* | | **azure.cognitiveservices_accounts.potentially_abusive_user_count**(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. | | **azure.cognitiveservices_accounts.safety_system_event**(gauge) | System event for risks & safety monitoring. You can add a filter or apply splitting by the following dimension: eventtype. | | **azure.cognitiveservices_accounts.call_count_for_image_moderation**(count) | Number of calls for image moderation. | | **azure.cognitiveservices_accounts.call_count_for_text_moderation**(count) | Number of calls for text moderation. | | **azure.cognitiveservices_accounts.job_duration_preview**(gauge) | Note: this value depends heavily on the input size, number of documents and task's complexity. This is an aggregate value across all job tasks.*Shown as millisecond* | | **azure.cognitiveservices_accounts.model_availability_rate**(gauge) | Availability percentage with the following calculation: (total calls - server errors)/total calls. Server errors include any HTTP responses >=500.*Shown as percent* | | **azure.cognitiveservices_accounts.model_requests**(count) | Number of calls made to the model API over a period of time. Applies to PTU, ptu-managed and pay-as-you-go deployments. | | **azure.cognitiveservices_accounts.time_between_tokens**(gauge) | For streaming requests; model token generation rate, measured in milliseconds. Applies to PTU and ptu-managed deployments.*Shown as millisecond* | | **azure.cognitiveservices_accounts.input_tokens**(count) | Number of prompt tokens processed (input) on a model. Applies to PTU, ptu-managed and pay-as-you-go deployments. | | **azure.cognitiveservices_accounts.output_tokens**(count) | Number of tokens generated (output) from an openai model. Applies to PTU, ptu-managed and pay-as-you-go deployments. | | **azure.cognitiveservices_accounts.provisioned_utilization**(gauge) | 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.*Shown as percent* | | **azure.cognitiveservices_accounts.total_tokens**(count) | Number of inference tokens processed on a model. Calculated as prompt tokens (input) plus generated tokens (output). Applies to PTU, ptu-managed and pay-as-you-go deployments. | | **azure.cognitiveservices_accounts.audio_seconds_batch_transcribed**(count) | Batch number of seconds transcribed. | | **azure.cognitiveservices_accounts.audio_seconds_batch_whisper_transcribed**(count) | Batch whisper number of seconds transcribed. | | **azure.cognitiveservices_accounts.audio_seconds_fast_transcribed**(count) | Fast number of seconds transcribed. | | **azure.cognitiveservices_accounts.audio_seconds_fast_whisper_transcribed**(count) | Fast whisper number of seconds transcribed. | | **azure.cognitiveservices_accounts.avatar_model_hosting_seconds**(count) | Number of seconds. | | **azure.cognitiveservices_accounts.avatar_model_training_seconds**(count) | Number of seconds. | | **azure.cognitiveservices_accounts.video_seconds_synthesized**(count) | Number of seconds synthesized. | | **azure.cognitiveservices_accounts.document_sync_characters_translated**(count) | Number of characters in document translation (synchronous) request. | | **azure.cognitiveservices_accounts.document_sync_custom_characters_translated**(count) | Number of characters in custom document translation (synchronous) request. | | **azure.cognitiveservices_accounts.translator_pro_app_seconds**(count) | Number of seconds of translator pro app usage.*Shown as second* | ### Events{% #events %} The Azure Cognitive Services integration does not include any events. ### Service Checks{% #service-checks %} The Azure Cognitive Services integration does not include any service checks. ## Troubleshooting{% #troubleshooting %} Need help? Contact [Datadog support](https://docs.datadoghq.com/help/).