The time range used for the LLM should be locked to the start/end of the investigation. TBD exactly how these time ranges relate, but the existing implementation just always uses "now" for the end of the LLM's target range, which causes the data volume to continually increase.
The time range used for the LLM should be locked to the start/end of the investigation. TBD exactly how these time ranges relate, but the existing implementation just always uses "now" for the end of the LLM's target range, which causes the data volume to continually increase.