README.md

Consumer service

Requirements

Java 11 is required to build and run this application. At runtime, the service relies on Kafka/Redpanda.
A simpler way to set up these dependencies would be to run the already supplied docker-compose. For further info please refer to solution README.

How to build

To build the so-called fat jar artifact, run ./mvnw clean package

Execution

To run the service, just type java -jar target/consumer-0.0.1-SNAPSHOT.jar

How it works

This service exposes an endpoint able to receive a batch of sensor data. The endpoint exposed can be called like this:

curl -L -X POST 'http://localhost:18888/sensor' \
-H 'Content-Type: application/json' \
--data-raw '[
    {
        "sensor_id": 1,
        "value": 100,
        "timestamp": 1611165641876
    },
    {
        "sensor_id": 1,
        "value": 110,
        "timestamp": 1611165641877
    },
    {
        "sensor_id": 1,
        "value": 120,
        "timestamp": 1611165641878
    },
    {
        "sensor_id": 1,
        "value": 10,
        "timestamp": 1611165641879
    },
    {
        "sensor_id": 1,
        "value": 20,
        "timestamp": 1611165641880
    }
]'

Each batch of sensor data will be then sent to a message queue called ingestion to supply data to the extractor service.

Implementation considerations

To quickly enable development of features such as REST endpoints and message production, the service relies on Spring Boot plus any needed dependency. In order to provide a way to eventually interchange message broker, a SensorDataSender interface has been provided.
For example, to switch from Kafka to RabbitMQ one should implement a RabbitMQ sender using the aforementioned interface - in addition to add the required package dependencies.