Collects the LLM inference calls made during a single operation, for display in a debug session. Holds at most the most recent 1000 inferences, discarding the oldest once that limit is reached, and is attached to the operation's captured OperationDetails when a debug session is capturing LLM activity. Written to the debug session's log as the llm-inference-stats element.

Group: Debugging


Properties

PropertyReturnsDescription
inferencesList<LlmInferenceInfo>The LLM inference calls recorded for this operation, in the order they were made.
totalDurationMillislongTotal time spent on the recorded inferences that have a known duration.

Methods

getInferences()

Returns: List<LlmInferenceInfo>

The LLM inference calls recorded for this operation, in the order they were made.

getTotalDurationMillis()

Returns: long

Total time spent on the recorded inferences that have a known duration.

To get full access to the Kademi Hub existing customers can login here, or new customers can register here.