To configure access control for Spark data, Zeppelin must be running as an end user ("identity propagation"). Zeppelin implements access control using Livy. When identity propagation is enabled via Livy, data access is controlled by the type of data source being accessed. For example, when you access HDFS data, access is controlled by HDFS permissions.

To configure access control for Hive data, use the JDBC interpreter.

To configure access control for the Spark shell, define permissions for end users running the shell.