I am reading about filebeat in ELK and because i am newbie, my question is does spooler also keep the actual data that harvester read on the disk or is it only about keeping the state of how much does harvester read data from input or send data to output?