Azure OpenAI "There was an error generating a response." or "An error occurred."

Chirag Joshi 10

I have deployed an Azure OpenAI chat using GPT 3.5 Turbo. The chat seems to be working for a while, but at a certain point, guaranteed, I get the following errors in the chat:

Error :

There was an error generating a response. Chat history can't be saved at this time. If the problem persists, please contact the site administrator.

Screenshot 2024-01-22 at 11.49.32 AM

I have used Sweden Central region.

Please help as we are paying for Azure OpenAI but without it actually working correctly.

navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-22T07:28:35.0533333+00:00
@Chirag Joshi Welcome to Microsoft Q&A Forum, Thank you for posting your query here!
.

Does your Azure OpenAi service work fine from the Open AI Studio --> Chat playground ?

Could you please click F12 button while accessing your webapp and check in the browser console tab? Check if there are any errors/exceptions ?
Edge:

Chrome:

Did you try re-deploying the app again and check if that helps ?

.

Awaiting your reply.
Chirag Joshi 10 Reputation points

2024-01-22T08:21:36.8733333+00:00
Question : Does your Azure OpenAi service work fine from the Open AI Studio --> Chat playground ?

Answer : Yes it works completely in the playground

Question : Check if there are any errors/exceptions ?

Answer :
navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-22T08:27:13.0266667+00:00

@Chirag Joshi Thanks for sharing the details. Could you please collect the browser HAR traces while reproducing this issue again and share the trace with me ?

https://learn.microsoft.com/en-us/azure/azure-portal/capture-browser-trace
Chirag Joshi 10 Reputation points

2024-01-22T09:42:26.8533333+00:00

It seems difficult to do following your documentation
navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-22T09:58:44.9333333+00:00
@Chirag Joshi Thanks for getting back. If you are facing challenges in collecting HAR traces, please collect fiddler trace and share it here.

Download and install fiddler

Launch fiddler. Click Tools > Options > HTTPS.

3.Select the Decrypt HTTPS Traffic box and click on subsequent Ok button pop-ups.

4.Then try to reproduce the issue.

5.Then save it by going to File-->Save All Sessions and share the traces here.
Chirag Joshi 10 Reputation points

2024-01-22T10:09:27.3+00:00

Screenshot of browser HAR.

Documentation to follow for get browser HAR : https://learn.microsoft.com/en-us/azure/azure-portal/capture-browser-trace
navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-22T10:29:02.69+00:00

@Chirag Joshi This screenshot will not help in troubleshooting this. It seems like you didn't collect it in the right way. To keep it simple, could you please collect the fiddler traces as mentioned in my above comment ?
Chirag Joshi 10 Reputation points

2024-01-22T11:21:17.8133333+00:00

Can we arrange a meeting?
navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-22T11:27:01.6833333+00:00

@Chirag Joshi Sure. We would be happy to have a Support professional to get in touch with you and assist you further in troubleshooting this issue. Could you please let me know if you have a Microsoft support plan ? If you don't have one, I can enable a one-time free support ticket for your subscription.
.
Please share the below details over email to azcommunity[at]microsoft[dot]com with subject as "Attn: NAVBA " , so that I can enable one-time free support ticket creation:

a. Subscription ID.
Markus Gutmann 0 Reputation points

2024-01-25T20:21:16.83+00:00

Did you solve this? I have the same issue

1 answer

Chirag Joshi 10 Reputation points

2024-01-26T03:12:17.78+00:00
That issue has been resolved but there is another issue. The issue is that when we give prompt our fine tuned model for the first time it generates a response but after again When I give a new prompt I get the following error.

Error Requests to the ChatCompletions_Create Operation under Azure OpenAI API version 2023-07-01-preview have exceeded call rate limit of your current OpenAI S0 pricing tier. Please retry after 8 seconds. Please go here: https://aka.ms/oai/quotaincrease if you would like to further increase the default rate limit. | Apim-request-id: 79d3e32e-a9e0-******

We have to solve this issue at any cost
Please sign in to rate this answer.
navba-MSFT 17,980 Reputation points Microsoft Employee

2024-01-26T05:40:21.9166667+00:00

@Chirag Joshi Thanks for getting back.
.
Please navigate to the Azure AI studio select the concerned Subscription and the Region to check the quota limit and the usage. This will help us in validating if you have exhausted the quota limits.

Note: If you have completed exhausted your quota and want to increase the quota limit then you need to fill this form and wait for the approval: https://aka.ms/oai/quotaincrease
.
The error message clearly mentions the exceeding of rate limit 10 requests per second (RPM): https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/quota?tabs=rest#understanding-rate-limits. Also ensure you are well within this limit. Go to https://aka.ms/oai/quotaincrease as indicated, if you need to increase this limit.

Markus Gutmann 0 Reputation points

2024-01-29T07:58:40.5266667+00:00

And how did you resolve the first issue?

Dennis Hasselgaard Larsen 5 Reputation points

2024-02-28T11:50:43.6066667+00:00

How did you fix the first problem?? An error occurred. Answers can't be saved at this time. If the problem persists, please contact the site administrator. I have the same error and would like to know how to fix this. How did you troubleshoot this

Prasanta Bhowmik 0 Reputation points

2024-05-08T13:01:31.9033333+00:00

ow did you fix the first problem?? An error occurred. Answers can't be saved at this time. If the problem persists, please contact the site administrator. I have the same error and would like to know how to fix this.
Sign in to comment

Share via

Azure OpenAI "There was an error generating a response." or "An error occurred."

1 answer