We need to have logging and monitoring or run-time telemetry logging for Fabric data agents. This will enable us to understand the question, responses provided by the agent and also the accuracy score that shows how to improvise the prompt accordingly.
There's Fabric SDK that will help us to just know the ping the question and get the response - but, it would be great to have run-time telemetry logging enabled somewhere for Fabric Data Agents.
1 Comment
PeterDanielsAdvocate II
Being able to see and analyze the time spent in various processing steps would be very helpful. Currently, we have to manually get a feeling for where time is spent.
