Azure language Text Analytics extract_key_phrases taking a long time to run

Question

Azure language Text Analytics extract_key_phrases taking a long time to run

PhuongN 0

I'm using the TextAnalytics in region Central US, for testing the sample code after deploying a resource with key and endpoint, but it is taking a very long time to run, other service works, such as computer vision service. Any issues with this service?

def authenticate_client():
    ta_credential = AzureKeyCredential(key)
    text_analytics_client = TextAnalyticsClient(
            endpoint=endpoint, 
            credential=ta_credential)
    return text_analytics_client

client = authenticate_client()

def key_phrase_extraction_example(client):

    try:
        documents = ["Dr. Smith has a very modern medical office, and she has great staff."]

        response = client.extract_key_phrases(documents = documents)[0]

        if not response.is_error:
            print("\tKey Phrases:")
            for phrase in response.key_phrases:
                print("\t\t", phrase)
        else:
            print(response.id, response.error)

    except Exception as err:
        print("Encountered exception. {}".format(err))
        
key_phrase_extraction_example(client)

Saideep Anchuri 6,620 Reputation points Microsoft External Staff

2025-05-07T01:41:14.94+00:00

Hi PhuongN

If you're experiencing slower performance with Azure Text Analytics compared to other services like Computer Vision, several factors could be contributing to the issue. One possible cause is region-specific performance variations, as the Central US region might be under heavy load or experiencing temporary issues. You can check the Azure Status Page for any ongoing incidents at https://learn.microsoft.com/en-us/azure/service-health/azure-status-overview.

Additionally, network latency might be affecting the response time of the Text Analytics service, and tools like ping or traceroute can help diagnose this. More details on network latency troubleshooting can be found at https://learn.microsoft.com/en-us/azure/network-watcher/connection-troubleshoot-overview. Optimizing your resource configuration can significantly improve performance. Ensure your resource settings are aligned with best practices by reviewing the Azure Text Analytics Documentation at Azure Text Analytics. Moreover, large input data sizes can slow down processing. Consider using Azure Blob Storage to host and manage large datasets efficiently, as explained in Azure Blob Storage. If your application is making high-frequency API calls, it may be subject to throttling, limiting throughput.

Understanding Azure API throttling can help manage request rates effectively, and more details can be found at Azure API Throttling. To troubleshoot, start by verifying the resource status in the Azure portal, checking for errors or warnings. Testing with different input data sizes can help identify if the issue is data-related. Conduct network latency checks to compare the response times between Text Analytics and other Azure services. Implementing parallel API calls and adjusting call frequency can smooth out traffic spikes, enhancing performance. Enabling diagnostic logging provides insights into recurring issues—guidance on Azure logs is available at Azure Diagnostic Logging. If performance remains inconsistent, deploying the resource in a different Azure region can help determine if the problem is location-specific. To maintain optimal performance, always use the latest version of the Text Analytics SDK, as outlined in Azure SDK Updates. Implement error handling and retry mechanisms, following best practices available at Azure Retry Patterns. Hosting large text datasets on Azure Blob Storage can further reduce latency and improve processing efficiency. Lastly, monitor API usage to avoid throttling by adjusting rate limits appropriately, as detailed at Azure API Limits.

Thank You.

Share via

Azure language Text Analytics extract_key_phrases taking a long time to run

Your answer