Deploying a Load-Balanced Data Pipeline

Deploying a Load-Balanced Data Pipeline

In this tutorial you install and deploy a containerized Ruby on Rails app named Tweeter. Tweeter is an app similar to Twitter that you can use to post 140-character messages to the internet. Then, you use Zeppelin to perform real-time analytics on the data created by Tweeter.

This tutorial demonstrates how you can build a complete IoT pipeline on DC/OS in about 15 minutes! You will learn:

  • How to install DC/OS services.
  • How to add apps to DC/OS Marathon.
  • How to route public traffic to the private application with Marathon-LB.
  • How your apps are discovered.
  • How to scale your apps.

This tutorial uses DC/OS to launch and deploy these microservices to your cluster:


The Cassandra database is used on the backend to store the Tweeter app data.


The Kafka publish-subscribe message service receives tweets from Cassandra and routes them to Zeppelin for real-time analytics.


Marathon-LB is an HAProxy based load balancer for Marathon only. It is useful when you require external routing or layer 7 load balancing features.


Zeppelin is an interactive analytics notebook that works with DC/OS Spark on the backend to enable interactive analytics and visualization. Because it’s possible for Spark and Zeppelin to consume all of your cluster resources, you must specify a maximum number of cores for the Zeppelin service.


Tweeter stores tweets in the DC/OS Cassandra service, streams tweets to the DC/OS Kafka service in real-time, and performs real-time analytics with the DC/OS Spark and Zeppelin services.


Install the DC/OS services you’ll need

From the DC/OS web interface Universe tab, install Cassandra, Kafka, Marathon-LB, and Zeppelin.

Tip: You can also install DC/OS packages from the DC/OS CLI with the dcos package install command.

  1. Find the cassandra package and click the Install Package button and accept the default installation. Cassandra will spin up to at least 3 nodes.
  2. Find the kafka package and click the Install Package button and accept the default installation. Kafka will spin up 3 brokers.
  3. Find the marathon-lb package and click the Install Package button and accept the default installation.
  4. Install Zeppelin.
    1. Find the zeppelin package and click the Install Package button and choose the Advanced Installation option.
    2. Click the spark tab and set cores_max to 8.
    3. Click Review and Install and complete your installation.
  5. Monitor the Services tab to watch as your microservices are deployed on DC/OS. You will see the Health status go from Idle to Unhealthy, and finally to Healthy as the nodes come online. This may take several minutes.

    Tip: It can take up to 10 minutes for Cassandra to initialize with DC/OS because of race conditions.

    Deployed services

Deploy the containerized app

In this step you deploy the containerized Tweeter app to a public node.

  1. Clone the Tweeter GitHub repository to your local directory.

    git clone
  2. Add the HAPROXY_0_VHOST label to the 1.8/tweeter.json Marathon app definition file. HAPROXY_0_VHOST exposes Nginx on the external load balancer with a virtual host. The HAPROXY_0_VHOST value is the hostname of your public agent node.

    Important: You must remove the leading http:// and the trailing /.

      "labels": {
        "HAPROXY_GROUP": "external",
        "HAPROXY_0_VHOST": "<Master-Public-IP>"

    For example, if you are using AWS, this is your public ELB hostname. It should look similar to this:

      "labels": {
        "HAPROXY_GROUP": "external",
        "HAPROXY_0_VHOST": ""
  3. Install and deploy Tweeter with this command.

    dcos marathon app add 1.8/tweeter.json

    Tip: The instances parameter in tweeter.json specifies the number of app instances. Use the following command to scale your app up or down:

    dcos marathon app update tweeter instances=<number_of_desired_instances>

    The service talks to Cassandra via node-0.cassandra.mesos:9042, and Kafka via broker-0.kafka.mesos:9557 in this example. Traffic is routed via the Marathon-LB (Marathon-LB) because you added the HAPROXY_0_VHOST tag on the tweeter.json definition.

  4. Go to the DC/OS web interface to verify your app is up and healthy. Then, navigate to public agent node to see the Tweeter UI and post a Tweet.


Post 100K Tweets

Use the 1.8/post-tweets.json app a large number of Shakespeare tweets from a file:

    dcos marathon app add 1.8/post-tweets.json

The app will post more than 100k tweets one by one, so you’ll see them coming in steadily when you refresh the page. Click the Network tab in the DC/OS web interface to see the load balancing in action.

The post-tweets app works by streaming to the VIP This address is declared in the cmd parameter of the post-tweets.json app definition. The app uses the service discovery and load balancer service that is installed on every DC/OS node. You can see the Tweeter app defined with this VIP in the json definition under VIP_0.

Add Streaming Analytics

Next, you’ll perform real-time analytics on the stream of tweets coming in from Kafka.

  1. Navigate to Zeppelin at https://<master_ip>/service/zeppelin/, click Import Note and import 1.8/tweeter-analytics.json. Zeppelin is preconfigured to execute Spark jobs on the DC/OS cluster, so there is no further configuration or setup required. Be sure to use https://, not http://.

    Tip: Your master IP address is the URL of the DC/OS web interface.

  2. Navigate to Notebook > Tweeter Analytics.

  3. Run the Load Dependencies step to load the required libraries into Zeppelin.

  4. Run the Spark Streaming step, which reads the tweet stream from ZooKeeper and puts them into a temporary table that can be queried using SparkSQL.

  5. Run the Top Tweeters SQL query, which counts the number of tweets per user using the table created in the previous step. The table updates continuously as new tweets come in, so re-running the query will produce a different result every time.

Top Tweeters