Rework kafka-avro example: drop Confluent Schema Registry, fix OOM

[Croway]

Summary

The kafka-avro example was broken in several ways:

• CachedSchemaRegistryClient(schemas, Integer.MAX_VALUE) in both the custom serializer and deserializer caused an OOM crash on startup, regardless of whether a schema registry was reachable.
• Real Avro object creation was commented out, so the example actually sent a plain String, not Avro.
• The avro-maven-plugin codegen was commented out in the pom.
• The Confluent schema registry URL was hardcoded.
• The README documented two contradictory setup paths (manual Confluent vs. camel infra run kafka) - neither worked alone, since infra kafka has no schema registry.

This rewrites the example to use Camel's native AvroDataFormat instead of Confluent's schema-registry-backed serializers. This needs only a plain Kafka broker (no schema registry at all), which removes the OOM-causing code outright and lets the README document a single, working setup path via camel infra run kafka.

Also fixes two latent bugs found while verifying end-to-end:

• KafkaConstants.KAFKA_RECORD_META needed recordMetadata=true on the producer endpoint to actually populate (was silently null), so the previously-unwired KafkaAvroProcessor is now wired into the producer route.
• consumer.group was defined in application.properties but never referenced in the consumer endpoint URI.

Test plan

• mvn clean package (checkstyle/RAT included) passes
• Started a broker with camel infra run kafka, ran mvn spring-boot:run, and confirmed the producer builds and sends real Employee Avro records, the consumer unmarshals and logs them back, and record metadata logging works with no errors/OOM.