Kafka 129 sits between BookNest's operational systems and its analytics (BookNest's Data Platform). The chapter uses three topics, filled with generated sample data from JSON, Columnar and Binary Formats's generator (seed 7):
| Topic | Key | Holds | Cleanup |
|---|---|---|---|
| booknest.order-events | order_id | 390,737 lifecycle events | Delete by time |
| booknest.customers | customer_id | Latest profile per customer | Compaction |
| booknest.catalog | book id | Latest metadata per book (XML and Its Toolchain) | Compaction |
The events carry 2025 times, so retention hinges on the record timestamp: The Kafka Log Under the Hood stamps records with the event time and sets retention.ms=-1, while the shop's producer stamps the send time and keeps 7 days (Producing Order Events). Kafka Connect 129 loads the events into PostgreSQL 1,289 (Kafka Connect); Kafka Streams 129 and Flink 129 compute live sales per genre (Kafka Streams and ksqlDB and Apache Flink); the lakehouse lands them in Iceberg 129 tables (Streaming into the Lakehouse); and Debezium 317,608 's change data capture from PostgreSQL joins them in Orchestration and Pipelines.