Skip to main content

Integrating with Confluent Cloud

DataHub provides the ability to easily leverage Confluent Cloud as your Kafka provider. To do so, you'll need to configure DataHub to talk to a broker and schema registry hosted by Confluent.

Doing this is a matter of configuring the Kafka Producer and Consumers used by DataHub correctly. There are 2 places where Kafka configuration should be provided: the metadata server (GMS) and the frontend server (datahub-frontend). Follow the steps below to configure these components for your deployment.

Step 1: Create topics in Confluent Control Center#

First, you'll need to create 4 new topics in the Confluent Control Center. By default they have the following names:

  1. MetadataChangeEvent_v4: Metadata change proposal messages
  2. MetadataAuditEvent_v4: Metadata change log messages
  3. FailedMetadataChangeEvent_v4: Failed to process #1 event
  4. DataHubUsageEvent_v1: User behavior tracking event for UI

To do so, navigate to your Cluster and click "Create Topic". Feel free to tweak the default topic configurations to match your preferences.


Step 2: Configure DataHub Container to use Confluent Cloud Topics#

Docker Compose#

If you are deploying DataHub via docker compose, enabling connection to Confluent is a matter of a) creating topics in the Confluent Control Center and b) changing the default container environment variables.

First, configure GMS to connect to Confluent Cloud by changing docker/gms/env/docker.env:

Next, configure datahub-frontend to connect to Confluent Cloud by changing docker/datahub-frontend/env/docker.env:

Note that this step is only required if DATAHUB_ANALYTICS_ENABLED is not set to false.

If you're deploying with Docker Compose, you do not need to deploy the Zookeeper, Kafka Broker, or Schema Registry containers that ship by default.


If you're deploying to K8s using Helm, you can simply change the datahub-helm values.yml to point to Confluent Cloud and disable some default containers:

First, disable the cp-schema-registry service:

cp-schema-registry:  enabled: false 

Next, disable the kafkaSetupJob service:

kafkaSetupJob:    enabled: false

Then, update the kafka configurations to point to your Confluent Cloud broker and schema registry instance, along with the topics you've created in Step 1:

kafka:      bootstrap:        server:      schemaregistry:        url:

Next, you'll want to create 2 new Kubernetes secrets, one for the JaaS configuration which contains the username and password for Confluent, and another for the user info used for connecting to the schema registry. You'll find the values for each within the Confluent Control Center. Specifically, select "Clients" -> "Configure new Java Client". You should see a page like the following:


You'll want to generate both a Kafka Cluster API Key & a Schema Registry key. Once you do so,you should see the config automatically populate with your new secrets:


You'll need to copy the values of sasl.jaas.config and for the next step.

The next step is to create K8s secrets containing the config values you've just generated. Specifically, you'll run the following commands:

kubectl create secret generic confluent-secrets --from-literal=sasl_jaas_config="<your-sasl.jaas.config>"kubectl create secret generic confluent-secrets --from-literal=basic_auth_user_info="<>"

With your config values substituted as appropriate. For example, in our case we'd run:

kubectl create secret generic confluent-secrets --from-literal=sasl_jaas_config="   required username='XFA45EL1QFUQP4PA' password='ltyf96EvR1YYutsjLB3ZYfrk+yfCXD8sQHCE3EMp57A2jNs4RR7J1bU9k6lM6rU';"kubectl create secret generic confluent-secrets --from-literal=basic_auth_user_info="P2ETAN5QR2LCWL14:RTjqw7AfETDl0RZo/7R0123LhPYs2TGjFKmvMWUFnlJ3uKubFbB1Sfs7aOjjNi1m23"

Finally, we'll configure our containers to pick up the Confluent Kafka Configs by changing two config blocks in our values.yaml file. You should see these blocks commented at the bottom of the template. You'll want to uncomment them and set them to the following values:

credentialsAndCertsSecrets:  name: confluent-secrets  secureEnv:    sasl.jaas.config: sasl_jaas_config basic_auth_user_info

springKafkaConfigurationOverrides:      security.protocol: SASL      sasl.mechanism: PLAIN      client.dns.lookup: use_all_dns_ips      basic.auth.credentials.source: USER_INFO

Then simply apply the updated values.yaml to your K8s cluster via kubectl apply.


Accepting contributions for a setup script compatible with Confluent Cloud!

Currently the kafka-setup-job container we ship with is only compatible with a distribution of Kafka wherein ZooKeeper is exposed and available. A version of the job using the Confluent CLI would be very useful for the broader community.