Why Kafka streams ?

Key points

Kafka stream is client side lib, it run where your code is running and not on the kafka agents nodes.
2 entities in kafka streams, streams and tables, 2 kinds of tables local and global.
Each of the entity can be created from the other and from kafka topics
Default storage is rocks db, there is a built-in in memory storage and many 3rd parties, storage can be queried directly
Streams can be manipulated using functional combinators like map, flatmap and more, each combinator create a new underline kafka topic that is automatically managed and reclaim by kafka streams.
Streams can join with table to produce stream
Streams can window and aggregated

The word count example, this is how it can be run

Run kafka

docker-compose up

Run the example app

./gradlew bootRun

Send inputs

docker-compose exec kafka bash
kafka-console-producer --bootstrap-server localhost:9092 --topic words
[type your lines here]

Interactive query

curl -s http://localhost:8080/api/all | jq

Get the values from the local storage.

WordCountTests

ServiceOpsProcessingTest

UserActivitySessionTests