Karapace

Karapace is a schema registry with the permissive Apache 2 license that we are using for the Datahub data catalog since the default confluent schema registry uses the more restrictive Confluent Community License.

See https://karapace.io/.

Administration

There is a karapace service which should provide the http endpoint at localhost:8081.

The commands to inspect the schemas are in the Karapace README.

The state is stored on a kafka topic called _schemas.

Instances

We have two instances of karapace.

• karapace1001 - This serves the production deployments of datahub
• karapace1002 - This serves the staging deployment of datahub

The instances are currently configured identically, except for the kafka cluster used to store the schemas. karapace1002 uses the kafka-test cluster, whereas karapace1001 uses the kafka-jumbo cluster.