You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
- [ ] bug report -> please search issues before submitting
- [ ] feature request
- [x ] documentation issue or request
- [ ] regression (a behavior that used to work and stopped in a new release)
Minimal steps to reproduce
We have azure open AI, search service and storage account deployed. The app is running as a kubernetes pod from an AKS cluster. All these services are in a same VNet/Subnet. Open AI and search uses system managed identity and respective accesses are set to them on respective resources
Initially everything was public and the operation worked fine with APIKey based operation.
Recently I started facing issue when I put Open AI, search and storage account behind Private endpoint. The PE is also under the same VNet.
Any log messages given by the failure
{"error": {"requestid": "1a19c986-ace5-47e6-a679-f54d8ca6b239", "code": 400, "message": "Invalid AzureCognitiveSearch configuration detected: Call to get Azure Search index failed. Check if you are using the correct Azure Search endpoint and index name. If you are using key based authentication, check if the admin key is correct. If you are using access token authentication or managed identity of Azure OpenAI, check if the Azure Search has enabled RBAC based authentication and if the user identity or Azure OpenAI managed identity has required role assignments to access Azure Search resource [https://aka.ms/aoaioydauthentication]. If the Azure Search resource has no public network access, make sure enable trusted service of Azure Search.\nAzure Search Error: 403, message='Server responded with status 403. Error message: ', url='https://xxxxxxx.search.windows.net//indexes/6b5f62e9-3785-48b1-b565-faa1dd8d9cfd?api-version=2024-03-01-preview'\nServer responded with status 403. Error message: "}}
Expected/desired behavior
The admin API key is the same in the KV
The search index is there and I can do the vectorization on the portal.
In idea scenario, there should not be an issue with the access.
Here is how the connection is setup
app --> Open AI --> search
I believe it's the second part of the setup above that is failing behind a PE
OS and Version?
Windows 7, 8 or 10. Linux (which distribution). macOS (Yosemite? El Capitan? Sierra?)
The app is running on a Linux based environment
Versions
Mention any other details that might be useful
Thanks! We'll be in touch soon.
The text was updated successfully, but these errors were encountered:
This issue is for a: (mark with an
x
)Minimal steps to reproduce
Initially everything was public and the operation worked fine with APIKey based operation.
Recently I started facing issue when I put Open AI, search and storage account behind Private endpoint. The PE is also under the same VNet.
Any log messages given by the failure
{"error": {"requestid": "1a19c986-ace5-47e6-a679-f54d8ca6b239", "code": 400, "message": "Invalid AzureCognitiveSearch configuration detected: Call to get Azure Search index failed. Check if you are using the correct Azure Search endpoint and index name. If you are using key based authentication, check if the admin key is correct. If you are using access token authentication or managed identity of Azure OpenAI, check if the Azure Search has enabled RBAC based authentication and if the user identity or Azure OpenAI managed identity has required role assignments to access Azure Search resource [https://aka.ms/aoaioydauthentication]. If the Azure Search resource has no public network access, make sure enable trusted service of Azure Search.\nAzure Search Error: 403, message='Server responded with status 403. Error message: ', url='https://xxxxxxx.search.windows.net//indexes/6b5f62e9-3785-48b1-b565-faa1dd8d9cfd?api-version=2024-03-01-preview'\nServer responded with status 403. Error message: "}}
Expected/desired behavior
The admin API key is the same in the KV
The search index is there and I can do the vectorization on the portal.
In idea scenario, there should not be an issue with the access.
Here is how the connection is setup
app --> Open AI --> search
I believe it's the second part of the setup above that is failing behind a PE
OS and Version?
The app is running on a Linux based environment
Versions
Mention any other details that might be useful
The text was updated successfully, but these errors were encountered: