i see kafka connect invented its own runtime data type system in org.apache.kafka.connect.data
however i struggle to understand how this is used. the payload in kafka is bytes. kafka does not carry any "schema" metadata. so how does connect know what the schema is of a ConnectRecord? if i write json data then perhaps i can see how a schema can be inferred from the data. is this what is happening? does this means the schema inference gets done for every json blob (which seems expensive)? thanks! koert
