I\'m using spark structured streaming kafka (2.4.0) to process some events and I saw that is spending allot of time on fetching data from kafka. I\'m wondering if is any way