Apache Flink - Casino Streaming

This project is an exemplary Apache Flink application, which presents stream processing in casino.

Prerequisites

• Maven
• Docker compose
• Flink 1.6+

Overview

This repository allows you to run an example casino stream processing on Flink's cluster. Architecture:

• Casino sites - an area that gathers several machines. Example: The Calypso, The Chariot
• Machines - physical machines that plays casino games. All machines sends events to the Kafka queue. Example: M0001, M0002, etc.
• Kafka - message queue
• Flink - stream processing engine
• Elasticsearch - storage for processed events
• Kibana - data visualization

Running locally

There is prepared the docker-compose.yaml file which allows you to run the following services:

• Elasticsearch - storage for processed events (sink)
• Kibana - data visualization
• Kafka - message queue
• Zookeeper - configuration service for Kafka

To run all services, execute the following command:

docker-compose up -d

To crate all necessary java libraries (Flink's job and casino simulation), execute the following command:

mvn package

To run Flink's local cluster, execute the following command in Flink's bin directory:

start-cluster

To deploy job, run the following command:

flink run -d core\target\casino-streaming-core-1.0-SNAPSHOT.jar

To run casino simulation, execute the following command:

java -jar simulation/target/casino-streaming-simulation-1.0-SNAPSHOT.jar

Finally, go to the Kibana page, open the Dashboard menu, click add a new dashboard, click add panel button and select all visualisations and saved searches.