Using Amazon EventBridge with Amazon Transcribe - Amazon Transcribe
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Using Amazon EventBridge with Amazon Transcribe

With Amazon EventBridge, you can respond to state changes in your Amazon Transcribe jobs by initiating events in other Amazon Web Services. When a transcription job changes state, EventBridge automatically sends an event to an event stream. You create rules that define the events that you want to monitor in the event stream and the action that EventBridge should take when those events occur. For example, routing the event to another service (or target), which can then take an action. You could, for example, configure a rule to route an event to an Amazon Lambda function when a transcription job has completed successfully. To define EventBridge rules, refer to the following sections.

You can receive notifications for events through multiple channels, including email, Amazon Chatbot chat notifications, or Amazon Console Mobile Application push notifications. You can also see notifications in the Console Notifications Center. If you want to set up notifications, you can use Amazon User Notifications. Amazon User Notifications supports aggregation, which can reduce the number of notifications you receive during specific events.

Defining EventBridge rules

To define EventBridge rules, use the Amazon Web Services Management Console. When you define a rule, use Amazon Transcribe as the service name. For an example of how to create an EventBridge rule, see Amazon EventBridge rules.

Before using EventBridge, note the following definitions:

  • Event–An event indicates a change in the state of one of your transcription jobs. For example, when the TranscriptionJobStatus of a job changes from IN_PROGRESS to COMPLETED.

  • Target–A target is another Amazon Web Service that processes an event. For example, Amazon Lambda or Amazon Simple Notification Service (Amazon SNS). A target receives events in JSON format.

  • Rule–A rule matches incoming events that you want EventBridge to watch for and routes them to a target or targets for processing. If a rule routes an event to multiple targets, all of the targets process the event in parallel. A rule can customize the JSON sent to the target.

Amazon EventBridge events are emitted on a best-effort basis. For more information about creating and managing events in EventBridge, see Amazon EventBridge events in the Amazon EventBridge User Guide.

The following is an example of an EventBridge rule for Amazon Transcribe that's initiated when a transcription job's status changes to COMPLETED or FAILED.

{ "source": [ "aws.transcribe" ], "detail-type": [ "Transcribe Job State Change" ], "detail": { "TranscriptionJobStatus": [ "COMPLETED", "FAILED" ] } }

The rule contains the following fields:

  • source–The source of the event. For Amazon Transcribe, this is always aws.transcribe.

  • detail-type–An identifier for the details of the event. For Amazon Transcribe, this is always Transcribe Job State Change.

  • detail–The new job status of the transcription job. In this example, the rule initiates an event when the job status changes to COMPLETED or FAILED.

Amazon Transcribe events

Amazon EventBridge logs several Amazon Transcribe events:

These events all contain the following shared fields:

  • version: The version of the event data. This value is always 0.

  • id: A unique identifier generated by EventBridge for the event.

  • detail-type: An identifier for the details of the event. For example, Transcribe Job State Change.

  • source: The source of the event. For Amazon Transcribe this is always aws.transcribe.

  • account: The Amazon Web Services account ID of the account that generated the API call.

  • time: The date and time the event is delivered.

  • region: The Amazon Web Services Region in which the request is made.

  • resources: The resources used by the API call. For Amazon Transcribe, this field is always empty.

  • detail: Additional details about the event.

    • FailureReason: This field is present if the state or status changes to FAILED, and describes the reason for the FAILED state or status.

    • Each event type has additional unique fields that are displayed under detail. These unique fields are defined in the following sections after each event example.

Transcription job events

When a job's state changes from IN_PROGRESS to COMPLETED or FAILED, Amazon Transcribe generates an event. To identify the job that changed state and initiate the event in your target, use the event's TranscriptionJobName field. An Amazon Transcribe event contains the following information. A FailureReason field is added under detail if your transcription job status is FAILED.

Note that this event applies only to the StartTranscriptionJob API operation.

{ "version": "0", "id": "event ID", "detail-type":"Transcribe Job State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "TranscriptionJobName": "my-first-transcription-job", "TranscriptionJobStatus": "COMPLETED" (or "FAILED") } }
  • TranscriptionJobName: The unique name you chose for your transcription job.

  • TranscriptionJobStatus : The status of the transcription job. This can be COMPLETED or FAILED.

Language identification events

When you enable automatic language identification, Amazon Transcribe generates an event when the language identification state is COMPLETED or FAILED. To identify the job that changed state and initiate the event in your target, use the event's JobName field. An Amazon Transcribe event contains the following information. A FailureReason field is added under detail if your language identification status is FAILED.

Note that this event applies only to the StartTranscriptionJob API operation when the LanguageIdSettings parameter is included.

{ "version": "0", "id": "event ID", "detail-type": "Language Identification State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "JobType": "TranscriptionJob", "JobName": "my-first-lang-id-job", "LanguageIdentificationStatus": "COMPLETED" (or "FAILED") } }
  • JobType: For transcription jobs, this value must be TranscriptionJob.

  • JobName: The unique name of your transcription job.

  • LanguageIdentificationStatus: The status of language identification in a transcription job. This can be COMPLETED or FAILED.

Call Analytics events

When a Call Analytics job state changes from IN_PROGRESS to COMPLETED or FAILED, Amazon Transcribe generates an event. To identify the Call Analytics job that changed state and initiates the event in your target, use the event's JobName field. An Amazon Transcribe event contains the following information. A FailureReason field is added under detail if your Call Analytics job status is FAILED.

Note that this event applies only to the StartCallAnalyticsJob API operation.

{ "version": "0", "id": "event ID", "detail-type": "Call Analytics Job State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "JobName": "my-first-analytics-job", "JobStatus": "COMPLETED" (or "FAILED") } }
  • JobName: The unique name of your Call Analytics transcription job.

  • JobStatus: The status of your Call Analytics transcription job. This can be either COMPLETED or FAILED.

Call Analytics post-call events

When a post-call analytics transcription changes state from IN_PROGRESS to COMPLETED or FAILED, Amazon Transcribe generates an event. To identify the Call Analytics post-call job that changed state and initiate the event in your target, use the event's StreamingSessionId field.

Note that this event applies only to the StartCallAnalyticsStreamTranscription API operation when the PostCallAnalyticsSettings parameter is included.

A COMPLETED event contains the following information:

{ "version": "0", "id": "event ID", "detail-type": "Call Analytics Post Call Job State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "StreamingSessionId": "session-id", "PostCallStatus": "COMPLETED", "Transcript": { "RedactedTranscriptFileUri": "s3://DOC-EXAMPLE-BUCKET/my-output-files/my-redacted-file.JSON", "TranscriptFileUri": "s3://DOC-EXAMPLE-BUCKET/my-output-files/my-file.JSON" }, "Media": { "MediaFileUri": "s3://DOC-EXAMPLE-BUCKET/my-output-files/my-redacted-file.WAV", "RedactedMediaFileUri": "s3://DOC-EXAMPLE-BUCKET/my-output-files/my-redacted-file.WAV" } } }

A FAILED event contains the following information:

{ "version": "0", "id": "event ID", "detail-type": "Call Analytics Post Call Job State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "StreamingSessionId": "session-id", "PostCallStatus": "FAILED" } }
  • StreamingSessionId: The identification number assigned to your real-time Call Analytics transcription request.

  • PostCallStatus: The status of your post-call Call Analytics transcription. This can be either COMPLETED or FAILED.

  • Transcript: The URI of your redacted and unredacted transcripts.

  • Media: The URI of your redacted and unredacted audio files.

Vocabulary events

When a custom vocabulary's state changes from PENDING to READY or FAILED, Amazon Transcribe generates an event. To identify the custom vocabulary that changed state and initiate the event in your target, use the event's VocabularyName field. An Amazon Transcribe event contains the following information. A FailureReason field is added under detail if your custom vocabulary state is FAILED.

Note that this event applies only to the CreateVocabulary API operation.

{ "version": "0", "id": "event ID", "detail-type": "Vocabulary State Change", "source": "aws.transcribe", "account": "111122223333", "time": "timestamp", "region": "us-west-2", "resources": [ ], "detail": { "VocabularyName": "unique-vocabulary-name", "VocabularyState": "READY" (or "FAILED") } }
  • VocabularyName: The unique name of your custom vocabulary.

  • VocabularyState: The processing state of your custom vocabulary. This can be READY or FAILED.