[promptflow.core._prompty_utils][ERROR] 500

Mohab Adnan 0

I made RAG as shown in :
https://learn.microsoft.com/en-us/azure/ai-studio/tutorials/copilot-sdk-create-resources?tabs=linux

but when i evaluate the model (Part 3 of the tutorial) i face this problem :
[2025-02-23 23:59:51 +0200][promptflow.core._prompty_utils][WARNING] - InternalServerError #3, but no Retry-After header, Back off 10 seconds for retry.

[2025-02-24 00:00:02 +0200][promptflow.core._prompty_utils][ERROR] - Exception occurs: InternalServerError: Error code: 500 - {'error': {'code': 'InternalServerError', 'message': 'Backen

Manas Mohanty 945 Microsoft Vendor

Hi Mohab Adnan

You can implement retry mechanism with a sleep time.

import time
import random
def retry_with_backoff(retries, func, *args, **kwargs):
    for i in range(retries):
        try:
            return func(*args, **kwargs) # your inference calls
        except InternalServerError as e:
            wait_time = (2 ** i) + random.uniform(0, 1)
            print(f"Retrying in {wait_time} seconds...")
            time.sleep(wait_time)
    raise Exception("Max retries exceeded")
# Example usage
result = retry_with_backoff(5, evaluate_model, model)

You can check other regions for same model or different version for same model or change deployment standard to "Global standard or Data zone standard" for more availability.

Could you tell the model, region and confirm whether you have correctly used endpoint URL and keys.

Thank you.

Manas Mohanty 945 Reputation points Microsoft Vendor

2025-02-27T12:10:42.1+00:00

Hi Mohab Adnan

We are waiting for your inputs. Hope the pointers were useful to you.

Thank you.

Share via

[promptflow.core._prompty_utils][ERROR] 500

Your answer