When Debezium connectors are deployed to Kafka Connect clusters, the captured change events are written directly to Kafka. Applications and services that want to consume these change events can do so by connecting to Kafka and using the KafkaConsumer. However, some applications will want to directly embed one or more Debezium connectors, in which case these applications will immediately consume the connectors' change events without using Kafka.

Debezium should provide a new API and implementation that allows applications consuming change events to use either or both of these techniques. Additional implementations for specialized environments (e.g., Spring Boot, Wildfly, etc.) could also be provided in the future.