Issue Querying Elasticsearch Logs with Azure OpenAI

Question

Hello, I am using the Elasticsearch (Preview) module in Azure OpenAI with the "Bring Your Own Data" service. I have successfully connected my Azure OpenAI instance to my Elasticsearch Cloud index, and I would like to query the log content of my index. However, even though the connection is established and my index is detected, all chatbot responses are: "The requested information was not found in the retrieved data. Please try a different query or topic." It seems unable to find any information, despite multiple attempts with different indexes. I have tried both keyword search and vector search, but without success. I don't understand what the issue is. Could you help me? Thanks in advance.

Answer 1

Hi AMROUN Lysa,

I understand that you're experiencing issues with querying your Elasticsearch logs through Azure OpenAI. Here are some steps can you follow.

• Ensure that the fields in your Elasticsearch index are correctly mapped. If the fields containing the log content are not properly defined, the model may not be able to retrieve the relevant information. You can customize the field mapping when adding your data source to ensure that the necessary fields are included.
• Since you mentioned trying both keyword and vector searches, ensure that you are using the correct search type based on your data. For keyword searches, make sure the terms you are querying exist in the indexed data. For vector searches, ensure that you have an appropriate embedding model deployed and that the vector embeddings are correctly set up.
• If the model responds with "The requested information isn't present in the retrieved documents," it may indicate that the queries are not matching any content in your index. You might want to try simpler queries or check if the data ingestion was successful and the logs are present in the index.
• If you are hitting any quota limitations, it may affect the ability to retrieve data. Ensure that you have not exceeded the limits for your service.
• Ensure that the storage account and Elasticsearch instance are accessible with the credentials you are using. If there are permissions issues, the model may not be able to access the data

Hope this helps. Do let us know if you any further queries.

---

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.