This section describes the Kafka agents. The agents are used to forward and collect messages using Kafka. Kafka is a cluster-based software that is executed outside of . Kafka uses a high throughput publish-subscriber messaging model. It persists all messages and by design, it decouples the publisher from the subscriber. That means that the forwarding agent will keep writing to the log even if the collection agents terminate.

The Kafka forwarding agent is listed among the processing agents in Desktop while the Kafka collection agent is listed among the collection agents.

Prerequisites

The reader of this document should be familiar with:

• Apache Kafka

The Kafka agents enable you to configure workflows with improved scalability and fault tolerance. As part of the data collection, data is written to Kafka to secure it from being lost if a failure occurs, and each topic can be set up to be replicated across several servers.

Open

Example of Kafka workflows

Scaling

Using Kafka provides the capability to scale as required. One of the many ways to scale a Kafka cluster is when you create your Kafka configuration. It is recommended that when creating your Kafka configuration, you consider how many partitions you may eventually require and add more than you may currently require,  as this will make it easier to scale up at a later stage.

Open

Example of Scaling a Kafka Cluster

You can also refer to http://kafka.apache.org for guidance on scaling using partitions.

For information on how to use the different tools delivered with Kafka, see http://kafka.apache.org.

Authentication with Kerberos and Encryption with SSL

You can configure the Kafka profile to use authentication via Kerberos and SSL encryption. Supported properties for this are available in the Producer and Consumer tabs in the Kafka profile. See https://www.cloudera.com/documentation/kafka/latest/topics/kafka_security.html for further information on how to configure this.