-
Notifications
You must be signed in to change notification settings - Fork 2.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Impossibe to Deploy a Web App in Azure OpenAI studio : An error message #1071
Comments
After conducting multiple tests, we've determined that the web app functions properly only when using the updated new interface (look), and your model, web app, and resource group are all located within the same region. |
I'm having trouble as well. Switched to new interface. Deployment worked. Says web app is running.
Can you please look at this? I guess I could deploy a standard plan. :( Logs in Azure shown below. 2024-08-30T21:49:47.832464576Z: [ERROR] WARNING:root:No datasource configuration found in the environment -- calls will be made to Azure OpenAI without grounding data. |
Same problem here, only the resource group isn't created in the same region, and I don't think I have permission to change it. |
Has anyone debugged the issue with deploying web app, i still have this problem "Web App deployment failed |
I'm having the same issue, very frustrating. |
Same issue, tried in both new and old interface and all objects are in the same region. |
I am having the same issue with new and old interface. |
I am running into the same issue. I tried with various locations, including EastUS or EastUS2, with combinations of RGs in different ones. Looks like something changed since last week as I was able to deploy. |
I could not take photos as this was on my work laptop but I was facing the same issue and was able to deploy by doing the following:
I was also able to verify from the environment variables of my Web App that the deployment was indeed updated (was upgrading from gpt3 to 4-mini) I have all my services in East US except my data source (using Azure AI search) which is in West US |
wow! The deployment worked using the postman trick with put/bearer token and modifying the "azureOpenAIEmbeddingEndpoint" : {} to "azureOpenAIEmbeddingEndpoint" : { "value":""} |
I tried a similar workaround as follows:
However, I received the following error: "error":{ I'm not sure which part of the settings is causing this issue. |
Did anyone try today? Is it fixed? |
unfortunately I have not seen this before - I would suggest checking the request body is complete |
Just checked again and failed with the same message "InvalidTemplate: Deployment template validation failed: 'The value for the template parameter 'azureOpenAIEmbeddingEndpoint' at line '1' and column '8426'"... |
Is this error likely to be resolved or is it worth just using the above workarounds going forward? Seems like an unnecessary amount of steps for a new process. |
There are other potential workarounds that bypass the AzureOAI Studio deployment template. I experimented with the RM template based deployment that is invoked with a click on the "Deploy to Azure" icon in the Readme. Unfortunately it has its own bugs: the temperature and top_p are typed int instead of float (mismatching the type in the Python code), and the UI customization is apparently out of scope. A documentation link is broken in the relevant section of the Readme. Besides, not as straightforward as the template sniffing and rectification approach discussed above. |
I can't imagine they won't fix it - probably they just don't know about it.. Not sure how to even report it lol |
If everyone can upvote this discussion it could help lol! need this fixed asap |
Hi all, the template validation error in the playground deployment flow is a known issue with a fix underway. Apologies for the inconvenience. In the meantime deploying from this repo with the 'deploy to azure' button or the azure CLI are possible workarounds. |
i used the postman trick and got this error ; |
Appears to be fixed. |
not fixed for me testing 5 minutes ago in the new Azure AI Studio Env. (All resources set to "West Europe") |
No change on my side, unfortunately. I tried with a few combinations of location, RG etc. My successful deployments in the past 10 days were all independent of the Studio. I also tried the az webapp CLI based deployment as workaround and it works. Not a direct replacement however, but offers more control. |
@jschmHoBa @gaborvar the fix is making its way to all environments. For now, I believe it is fixed in Azure OpenAI studio, if you use the "old view". |
Yes, it seems like the bug has been fixed, and the command modification workaround is no longer necessary. However, I am now directly getting the following error when using the deploy button. This indicates that while the workaround mentioned in my previous post was correct, there are still some issues in my environment. Any suggestions to fix this? SubscriptionIsOverQuotaForSku: This region has quota of 0 instances for your subscription. Try selecting different region or SKU. |
It works, thank you! (West Europe location, new OpenAI Studio) |
This issue is stale because it has been open for 30 days with no activity. |
Describe the bug
Impossible to deploy to web app, multiple users confirmed that they have the following error message while trying to deploy:
"InvalidTemplate: Deployment template validation failed: 'The value for the template parameter 'azureOpenAIEmbeddingEndpoint' at line '1' and column '4690' is not provided. Please see
https://aka.ms/arm-create-parameter-file
for usage details.'."
To Reproduce
Steps to reproduce the behavior in old or new look:
Expected behavior
It should deploy the app but in the old look, I have the error "InvalidTemplate: Deployment template validation failed: 'The value for the template parameter 'azureOpenAIEmbeddingEndpoint' at line '1' and column '4690' is not provided. Please see
https://aka.ms/arm-create-parameter-file
for usage details.'."
In the new look, I also have an error but it's not specified.
Screenshots
The text was updated successfully, but these errors were encountered: