HALO Jumping into Flink: Lessons Learned from Managing Real-Time Data

Welcome to the companion repository for the HALO Jumping into Flink meetup!

This repository contains a small demo showcasing a geofencing example built with Apache Flink.
It demonstrates how Flink can be used to process real-time spatial data streams and detect when entities enter or leave defined geographic zones.

---

Contents

• flink-job/ – Flink job implementation for the geofencing PoC
• kafka-producer/ – A simple kafka producer that simulates streaming location data
• Dockerfile – Container definition for running the Flink job
• docker-compose.yml – Simple setup to run the PoC locally with Flink services

---

Running the Example

You can start the entire demo locally with:

docker compose build
docker compose up

or

# Build Docker images
make build

# Start core stack (Kafka + Flink)
make up

# Wait for jobmanager & taskmanagers to be healthy, then start producer
make producer

# View logs for all services
make logs

# Stop everything
make down

# Clean up everything (containers, networks, volumes)
make clean

Dataset

The dataset used for this demo is publicly accessible and can be found here. It includes the real locations of bus stops for Carris Metropolitana in the Lisbon area, Portugal.

Image source can be found here.