/iot-browser

Primary LanguageJava

Simple application for saving IoT data from Kafka, saving data in database and querying data via API.

Application listens to 3 Kafka topics: OutsideTemperature, OutsideHumidity, WindSpeed. Contracts for topics can be found here

Application saves data from topics to the database. In-memory H2 database was used. Data structure is flat and no-relational, so NoSQL database can be used in a real use case.

Application exposed API endpoint is /readings/.
JSON must be sent as a body for the POST request. Contracts for request and response can be found here
Postman collection for API calls

Dependencies

The full dependency list can be seen in pom.xml Bellow listed the main set of dependencies

  • Java11
  • Maven
  • SpringBoot
  • SpringCloudStream
  • JUnit
  • Mockito
  • JBehave
  • Serenity
  • H2

Kafka cluster for the local running of application and functional testing created with docker-compose

Running

  1. Checkout repository
  2. Run docker-compose up -d to create Kafka and Redis containers
  3. Run mvn clean package

The easiest way is to run it inside IntelliJ IDEA. Steps :

  1. Run src/main/java/com/relay42/browser/IotBrowser.java

If there is no IDEA.

  1. Run java -jar target/iotbrowser-DEVELOP-SNAPSHOT.jar

Testing

Majority of classes are covered with unittests.

Functional flows are covered with functional tests. Serenity BDD approach was used to have tests as documentation. Stories can be found here

For almost real testing sends_random_kafka_messages.story was created. It sends 3 different messages every second in infinite loop. It is possible to run this story and make calls to application via Postman.

Known issues and possible improvements

  • Different types of devices can have the same id or can be members of the same group. There are 2 major ways to solve this problem:
    • send type of device in the request. So only reading from devices with particular type will be fetched.
    • put responsibility of providing unique ids and valid group ids to kafka messages producer (chosen)
  • Max, min, average, etc. can be done via database function(if engine supports of course). But I see it as coupling to DB engine. That's why I decided to calculate values in app from the list of models.