Supercharging Kafka Applications with Real-time Contextual Insights
summary
Demo shows Kafka topics mapped into Hazelcast via continuous SQL mapping and a Go CLI, running in Docker with a single-node demo cluster. SQL jobs generate or read Kafka events, join streaming orders with an in-memory IMAP to emit enriched records, and write results to a destination map so queries see persisted rows. A production pattern runs five parallel ML models inside the cluster with a majority vote for fraud decisions, yielding real-time, in‑memory enrichment while the customer waits.
AI-written summary · about
chapters (9)
- 0:00Introduction to real-time insights
- 2:31Hazelcast platform overview
- 5:00Demo setup and Kafka integration
- 9:00Reading Kafka stream with SQL
- 12:11Enriching Kafka data in real time
- 15:48Storing and querying enriched data
- 16:31Combining multiple Kafka streams
- 17:12Fraud detection use case
- 18:18Summary and closing
related talks
49mEnriching Kafka Applications With Real-time Contextual Data
30mFive Challenges in Real-time Stream Processing and Five Solutions
53mApache Kafka and KSQL in Action : Lets Build a Streaming Data Pipeline!
43mApache Kafka and KSQL in Action: Let’s Build a Streaming Data Pipeline!
01h 00mIn Depth Kafka Streams
36mInteractive Session: Running Flink SQL jobs in production
43mContinuous SQL with Kafka and Flink
46mFlink SQL in 2020: Time to show off!