Spring Kafka: Template, Listeners, and Config

KafkaTemplate and producers

kafkaTemplate.send(topic, key, value) uses configured ProducerFactory. DefaultKafkaProducerFactory shares one producer per config. VaultCommerce defines multiple templates only when serialization differs (Avro vs JSON).

@KafkaListener configuration

concurrency=4 spawns 4 consumer threads (≤ partitions). ackMode=MANUAL_IMMEDIATE commits after listener returns. groupId from property placeholder per environment. VaultCommerce uses @RetryableTopic on critical listeners.

VaultCommerce rollout checklist

Before promoting changes that touch the VaultCommerce order and payment event backbone, run the staging KRaft cluster (Kafka 3.7+, Schema Registry 7.x) through a 10k events/min soak test. Compare producer request latency p99 and consumer lag per group against the pre-deploy baseline. KafkaTemplate for send; @KafkaListener for consume. Document the change in the internal topic registry, attach Grafana screenshots to the change ticket, and keep an engineer on lag dashboards for 30 minutes after production rollout — roll back the service release before altering broker-level settings if lag or under-replicated partitions spike.

// VaultCommerce order-service — Spring Kafka 3.x
@Service
@RequiredArgsConstructor
public class OrderEventPublisher {
    private final KafkaTemplate<String, OrderPlaced> kafkaTemplate;

    public void publish(OrderPlaced event) {
        kafkaTemplate.send("vaultcommerce.orders.placed.v1", event.orderId(), event)
            .whenComplete((result, ex) -> {
                if (ex != null) throw new EventPublishException(event.orderId(), ex);
            });
    }
}

@KafkaListener(topics = "vaultcommerce.orders.placed.v1", groupId = "inventory-service", concurrency = "4")
public void handle(ConsumerRecord<String, OrderPlaced> record) {
    inventoryService.reserve(record.value());
}