Kafka Deserializers

Deserializers in Kafka

This passage explains the importance of using matching serializers and deserializers in Kafka.

When a producer sends data to Kafka, it must first serialize that data — convert it from an object (like a number or string) into a sequence of bytes that Kafka can store and transmit. On the consumer side, the same data must be deserialized — converted back from bytes into a usable form.

If the serializer and deserializer don’t match, the consumer will not be able to correctly interpret the data. For example, if you use IntSerializer to send integers but the consumer uses StringDeserializer, the consumer will misread the byte data and likely fail with an error.

As a result, developers must know which serializers were used for each Kafka topic and ensure that all consumers of that topic use the corresponding deserializers.

Using a Schema Registry with Avro serialization helps enforce this consistency. The Avro serializer and deserializer rely on a defined schema that ensures all data in a topic follows the same structure. If a producer tries to write incompatible data or if a consumer uses the wrong schema, Kafka will throw clear errors.

This makes debugging much easier and prevents data corruption or misinterpretation.

Custom Deserializers in Kafka

Check pg 184 for more details.

Using Avro Deserialization with Consumer

1 - Use KafkaAvroDeserializer to deserialize Avro messages 2 - To store schemas 3 - Define type for the record value
4 - Instance of customer