Aiven “rolling challenge” with Apache Kafka® and Apache Flink®

Welcome to Aiven’s “Rolling” challenge, an easy way for you to explore Aiven for Apache Kafka® and Aiven for Apache Flink®.

With the launch of Aiven for Apache Flink® we added a new way to manipulate your Apache Kafka® streaming data via SQL statements, providing the best combination of tools for real-time data transformation.

For this challenge, we’ll be using Aiven fake data generator on Docker to generate a series of symbols. The challenge consists of understanding the overall meaning of the symbols by transforming the original series of data with Apache Flink.

Let’s dive right in.

Instructions

The goal is to make sense of the incoming stream of data.

1. Create an Aiven free trial account: sign up for free.

2. Create an Aiven for Apache Kafka® and Aiven for Apache Flink® service

3. Set up an integration between the Aiven for Apache Kafka® and Apache Flink® services

4. Create a new Aiven authentication token

5. Clone the Aiven fake data generator on Docker with:

   git clone https://github.com/aiven/fake-data-producer-for-apache-kafka-docker
   

6. Copy the file conf/env.conf.sample to conf/env.conf and edit the following parameters

   Parameter Name	Parameter Value
   PROJECT_NAME	Name of the Aiven Project where the Aiven for Apache Kafka service is running
   SERVICE_NAME	Name of the Aiven for Apache Kafka service running
   TOPIC	Name of the Topic to write messages in. rolling for the challenge
   PARTITIONS	5
   REPLICATION	2
   NR_MESSAGES	0
   MAX_TIME	0
   SUBJECT	rolling
   USERNAME	Your Aiven account username
   TOKEN	Your Aiven account token
   PRIVATELINK	NO
   SECURITY	SSL

7. Build the Docker image

   docker build -t fake-data-producer-for-apache-kafka-docker .
   

8. Run the Docker image

   docker run fake-data-producer-for-apache-kafka-docker
   

9. Check the fake messages being produced by Docker

10. In the Aiven Console, navigate to the Aiven for Apache Flink service page

11. Play with the Aiven for Apache Flink Application tab and try to make sense of the data.

    Tip

    The source table can be mapped in Aiven for Apache Flink with the following SQL, using the rolling topic as source:

    CREATE TABLE ROLLING_IN(
        ts BIGINT,
        val string,
        ts_ltz AS TO_TIMESTAMP_LTZ(ts, 3),
        WATERMARK FOR ts_ltz AS ts_ltz - INTERVAL '10' SECOND
        )
    WITH (
        'connector' = 'kafka',
        'properties.bootstrap.servers' = '',
        'topic' = 'rolling',
        'value.format' = 'json',
        'scan.startup.mode' = 'earliest-offset'
        )
    

12. When you find the solution, email a screenshot to hacks@Aiven.io

Tips

Some tips that could help in solving the challenge:

• kcat is a tool to explore data in Apache Kafka topics, check the dedicate documentation to understand how to use it with Aiven for Apache Kafka

• jq is a helpful tool to parse JSON payloads, read the instructions on how to install and check the following useful flags:

  • -r retrieves the raw output

  • -j doesn’t create a new line for every message

  • -c shows data in compact view

• If you’re stuck with visualizing kcat consumer data with jq, check the -u flag as per dedicated example