My question: is it possible to implement
logic in the FlinkKafkaProducer to serialize message and store schema id in the Confluent Schema registry?
I don't think this is going to work with the current interface because creation and caching of the schema id in the Confluent Schema Registry is done with the help
of io.confluent.kafka.serializers.KafkaAvroSerializer.class and
all FlinkKafkaProducer constructors have either SerializationSchema or KeyedSerializationSchema (part of Flink's own serialization stack) as one of the parameters.
If my assumption is wrong, could you please provide details of implementation?
Thank you very much,