Coffeeshop Demo with Quarkus

This directory contains a set of demo around reactive in Quarkus with Kafka. It demonstrates the elasticity and resilience of the system.

Build

mvn clean package

Prerequisites

Install test HTTP client:

brew install httpie

Run Kafka with:

docker-compose up

In case of previous run, you can clean the state with

docker-compose down
docker-compose rm

Then, create the orders topic with ./create-topics-docker.sh

Run the demo

You need to run:

• the coffee shop service
• the HTTP barista
• the Kafka barista

Im 3 terminals:

cd coffeeshop-service
mvn compile quarkus:dev

cd barista-http
java -Dbarista.name=tom -jar target/barista-http-1.0-SNAPSHOT-runner.jar

cd barista-kafka
mvn compile quarkus:dev

Execute with HTTP

The first part of the demo shows HTTP interactions:

• Barista code: me.escoffier.quarkus.coffeeshop.BaristaResource
• CoffeeShop code: me.escoffier.quarkus.coffeeshop.CoffeeShopResource.http
• Generated client: me.escoffier.quarkus.coffeeshop.http.BaristaService

Order coffees with:

while [ true ]
do
http POST :8080/http product=latte name=clement
http POST :8080/http product=expresso name=neo
http POST :8080/http product=mocha name=flore
done

Stop the HTTP Barista, you can't order coffee anymore.

Execute with Kafka

• Barista code: me.escoffier.quarkus.coffeeshop.KafkaBarista: Read from orders, write to queue

• Bridge in the CoffeeShop: me.escoffier.quarkus.coffeeshop.messaging.CoffeeShopResource#messaging just enqueue the orders in a single thread (one counter)

• Get prepared beverages on me.escoffier.quarkus.coffeeshop.dashboard.BoardResource and send to SSE

• Open browser to http://localhost:8080/

• Order coffee with:

http POST :8080/messaging product=latte name=clement
http POST :8080/messaging product=expresso name=neo
http POST :8080/messaging product=mocha name=flore

Baristas do breaks

1. Stop the Kafka barista
2. Continue to enqueue order

http POST :8080/messaging product=frappuccino name=clement
http POST :8080/messaging product=chai name=neo
http POST :8080/messaging product=hot-chocolate name=flore

1. On the dashboard, the orders are in the "IN QUEUE" state
2. Restart the barista
3. They are processed

2 baristas are better

1. Start a second barista with:

java -Dquarkus.http.port=9095 -Dbarista.name=tom -jar target/barista-kafka-1.0-SNAPSHOT-runner.jar

1. Order more coffee

http POST :8080/messaging product=frappuccino name=clement
http POST :8080/messaging product=chai name=neo
http POST :8080/messaging product=hot-chocolate name=flore
http POST :8080/messaging product=latte name=clement
http POST :8080/messaging product=expresso name=neo
http POST :8080/messaging product=mocha name=flore

The dashboard shows that the load is dispatched among the baristas.