Zeppelin currently lacks features for data source metadata exploration. This is often a need for those building dashboards and composing queries. This is especially true for people working with large RDBMS. In most cases, the requisite functionality for these users is simple: see what tables and views are available and see what column metadata is there for a given table or view.
This is a modest early step in adding metadata support and focuses on the JDBC Interpreter. It creates a keyword, "explore" which can be used either in isolation for data source metadata or followed with the name of a database object for object metadata.
See a video example of this proposed feature here (https://s3.amazonaws.com/screenshots-mockups/embedvid.html).