Program & Containerize Your Kafka Producer/Consumer In Minutes

In this article, I’ll show how you can build your own Kafka producer/consumer in minutes.

We often need to be able to program our own producer/consumer for demos, functional testing, and sometimes even scale testing.

In this article, I’ve used a pip3 library called kafka-python in order to interact with a containerized Kafka cluster installed as part of this article.

As Kubernetes is a major factor in our infrastructure today, we’ll containerize our producer/consumer as well to be able to lift-and-shift our producer/consumer into our Kubernetes cluster.

So let’s go!


  • Podman installed on your computer (I’ve added the docker-podman package as well)

Note: The docker-podman package is an alias that gives the ability to use Docker commands although there is no Docker Engine underneath.

Setting Up A Containerized Kafka Cluster

In order to run our containerized Kafka cluster, We'll use single-node based services, so our Zookeeper and Broker will be installed on a single container.

Let’s start by creating a shared network for our Kafka cluster:

$ docker network create kafka

This will create an isolated network under a network namespace, that will allow our containers to share subnets, DNS records, etc.

Now we’ll initialize our Zookeeper service:

$ docker run -d --name zookeeper --net kafka -e ZOOKEEPER_CLIENT_PORT=2181 -e ZOOKEEPER_TICK_TIME=2000 -p 22181:2181 confluentinc/cp-zookeeper:latest

As you can see, we’ve run the Zookeeper container with a hardened name, and attached it to the previously created network. We've also exposed the Zookeeper port, which is 2181 to 22181 on the host, just in case we'll have to access it externally.

Great! now that we have Zookeeper running, Let's set up our Kafka broker, to start to write/read events:


As with Zookeeper, we gave our Broker container a name and attached it to the shared network, so that the cluster will be able to resolve the attached containers.

Some important configurations to notice:

  • KAFKA_ZOOKEEPER_CONNECT connects to the name that was given to our Zookeeper container, using the DNS record that was created automatically by our shared network
  • KAFKA_ADVERTISED_LISTENERS defines how should Kafka advertize its nodes, as we need it to listen to both our containerized network and the outside world, we'll tell Kafka that the broker name that will be returned for internal access (inside the containerized network) is kafka:9092. and for external network localhost:29092.

As you can see, we tell our Kafka broker that it'll have only one replica, and we map the 29092 port to the host itself, in case we'll need to access it externally.

Interacting With Our Kafka Topic

Let’s login to our Kafka cluster in order to create a topic:

$ docker exec -it kafka bash

Now let’s create a topic with Kafka's internal scripts:

$ kafka-topics --create --topic test --replication-factor 1 --partitions 1 -bootstrap-server kafka:9092Created topic test.

Great! let’s verify that everything we configured correctly:

$ kafka-topics --describe --topic test --bootstrap-server localhost:9092Topic: test	TopicId: AsiqI1nXTSO0xp8tYCi0Zw	PartitionCount: 1	ReplicationFactor: 1	Configs: 
Topic: test Partition: 0 Leader: 1 Replicas: 1 Isr: 1

We see that our topic was successfully created, Let’s test it.

We’ll use Kafka's internal scripts to interact with this topic.

First, we’ll send some messages using our Kafka producer:

$ kafka-console-producer --topic test --bootstrap-server kafka:9092>test
>This is test

Now, we’ll do the same to read the advertized messages:

$ kafka-console-consumer --topic test --bootstrap-server kafka:9092 --from-beginningtest
This is test

Great! but now we haven’t tested network access as this thing happened locally on our broker.

In order to test our created topic externally, we’ll have to use the external representation of our Kafka broker, which is localhost:29092.

In order to interact with our cluster, we’ll use a tool called kcat the simulates producer/consumer with an easy CLI:

$ kcat -b localhost:29092 -t test% Auto-selecting Consumer mode (use -P or -C to override)
This is test

As we’ve mapped the external port of our Kafka broker to port 29092 on the host, we're now able to connect to the cluster externally. In my case, It's localhost as I run it on my computer, but it could be any IP that you'll choose.

Note: If you plan on not using localhost, change KAFKA_ADVERTISED_LISTENERS to suit the external representation that you choose

Now let’s produce some events on the same topic using kcat:

$ kcat -b localhost:29092 -t test -PThis is test 2

Read the topic content once again:

$ kcat -b localhost:29092 -t test% Auto-selecting Consumer mode (use -P or -C to override)
This is test
This is test 2

Developing Our Own Kafka Producer/Consumer

Let’s start by writing the code that will eventually act as our Kafka producer:

This simple code will write 100 events to our Kafka cluster while sleeping for a second between iterations.

The script uses argparse to get the broker list, so as the wanted topic and prints to written messages to STDOUT.

In order to containerize this code snippet, we’ll have to use two more files.

The first one is the Dockerfile that holds the way our producer container image should be built:

The second one is the requirements.txt file that holds all of our dependencies:

Now let’s build our producer image:

$ docker build -t kafka-basic-producer .

Great, now we can test it in front of our containerized `Kafka cluster:

$ docker run --rm --name kafka-basic-producer --net kafka localhost/kafka-basic-producer -b 'kafka:9092' -t 'test'

As you can see, we’ve used kafka:9092 as the broker list, as this is the internal shared network. We've also used the test script that was created in previous stages.

Look at the container logs:

$ docker logs -f kafka-basic-producerNFO 2022-05-05 22:18:31,575 - Writing document {'number': 0} to Kafka cluster...
INFO 2022-05-05 22:18:32,579 - Writing document {'number': 1} to Kafka cluster...
INFO 2022-05-05 22:18:33,582 - Writing document {'number': 2} to Kafka cluster...
INFO 2022-05-05 22:18:34,585 - Writing document {'number': 3} to Kafka cluster...
INFO 2022-05-05 22:18:35,587 - Writing document {'number': 4} to Kafka cluster...
INFO 2022-05-05 22:18:36,589 - Writing document {'number': 5} to Kafka cluster...
INFO 2022-05-05 22:18:37,591 - Writing document {'number': 6} to Kafka cluster...
INFO 2022-05-05 22:18:38,594 - Writing document {'number': 7} to Kafka cluster...
INFO 2022-05-05 22:18:39,597 - Writing document {'number': 8} to Kafka cluster...
INFO 2022-05-05 22:18:40,599 - Writing document {'number': 9} to Kafka cluster...

Nice! we have our events written to our containerized Kafka cluster.

Let’s do the same with our consumer, Take a look at its code snippet:

Code looks quite the same, except that here we load our data from the cluster. Let’s build it the same way we did with our producer:

$ docker build -t kafka-basic-consumer .

Now let’s run it:

$ docker run --rm --name kafka-basic-consumer --net kafka localhost/kafka-basic-consumer -b 'kafka:9092' -t 'test'

When looking at the logs we see that now we have our messages read from the cluster:

$ docker logs -f kafka-basic-consumer INFO 2022-05-05 22:36:32,278 - Consuming document {'number': 0} from Kafka cluster...
INFO 2022-05-05 22:36:33,279 - Consuming document {'number': 1} from Kafka cluster...
INFO 2022-05-05 22:36:34,281 - Consuming document {'number': 2} from Kafka cluster...
INFO 2022-05-05 22:36:35,282 - Consuming document {'number': 3} from Kafka cluster...
INFO 2022-05-05 22:36:36,283 - Consuming document {'number': 4} from Kafka cluster...
INFO 2022-05-05 22:36:37,284 - Consuming document {'number': 5} from Kafka cluster...
INFO 2022-05-05 22:36:38,285 - Consuming document {'number': 6} from Kafka cluster...
INFO 2022-05-05 22:36:39,287 - Consuming document {'number': 7} from Kafka cluster...
INFO 2022-05-05 22:36:40,288 - Consuming document {'number': 8} from Kafka cluster...
INFO 2022-05-05 22:36:41,290 - Consuming document {'number': 9} from Kafka cluster...

Thank you very much for reading that far, Hope you’ve enjoyed our demo.

See ya next time!



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store