data flow is configured to read existing and new records.
If we configure it as only new records,its working fine by reading records only after the dataflow configuration.
the problem here is , if we configure read only new records and by what ever reason dataflow got stopped in the middle and at this point kafka topic has 500 records which are not processed by pega. When i restart dataflow, as its read only new records we are going to loose 500 records which are already in kafka topic.For this reason we choose for read Existing and new records,but with this configuration whenever data flow is restarted its processing all existing records again.
Pega platform - 8.1.4
In 7.3 we could see read from begining option in kafka stream data set which is missing in 8.1.4