This project has retired. For details please refer to its Attic page.

Analyze Big Data Platforms For Security and Performance

Data Classification Tutorial

Apache Eagle data classification feature provides the ability to classify data with different levels of sensitivity. Currently this feature is available ONLY for applications monitoring HDFS, Hive1 and HBase2. For example, HdfsAuditLog, HiveQueryLog and HBaseSecurityLog.

The main content of this page are

  • Cluster Connection
  • Data Classification

Cluster Connection

Here we give example configurations for HDFS, HBASE, and Hive. Suppose the cluster to monitor is Hortonwork sandbox. This configuration locates at the admin management page.

  • HDFS

    hdfs setup

    • Base case

      You may configure the default path for Apache Hadoop clients to connect remote hdfs namenode.

    • HA case

      Basically, you point your fs.defaultFS at your nameservice and let the client know how its configured (the backing namenodes) and how to fail over between them under the HA mode

    • Kerberos-secured cluster

      For Kerberos-secured cluster, you need to get a keytab file and the principal from your admin, and configure “eagle.keytab.file” and “eagle.kerberos.principal” to authenticate its access.


      If there is an exception about “invalid server principal name”, you may need to check the DNS resolver, or the data transfer , such as “”, “”, “dfs.trustedchannel.resolver.class”, “dfs.datatransfer.client.encrypt”.

  • Hive
    • Basic

  • HBase

    • Basic case

      You need to sett “hbase.zookeeper.quorum”:”localhost” property and “” property.
    • Kerberos-secured cluster

      According to your environment, you can add or remove some of the following properties. Here is the reference.

Any questions on the Kerberos configuration in Eagle, please first check FAQ

Data Classification

After the configuration is updated, we can go back to the classification page on Eagle UI. Here we take HdfsAuditLog as an example to explain how to classify data and how to monitor sensitive data in Eagle.

Part 1: Sensitivity Edit

  • add the sensitive mark to files/directories.

    • Basic: Label sensitivity files directly (recommended)

      HDFS classification HDFS classification HDFS classification

    • Advanced: Import json file/content

      HDFS classification HDFS classification HDFS classification

  • remove sensitive mark on files/directories

    • Basic: remove label directly

      HDFS classification HDFS classification

    • Advanced: delete lin batch

      HDFS classification

**Part 2: Monitor sensitive data

You can mark a particular folder/file as “PRIVATE”. Once you have this information you can create policies using this label.

For example: the following policy monitors all the operations to resources with sensitivity type “PRIVATE”.

sensitivity type policy


  1. All mentions of “hive” on this page represent Apache Hive. 

  2. All mentions of “hbase” on this page represent Apache HBase. 

Copyright © 2015 The Apache Software Foundation, Licensed under the Apache License, Version 2.0.
Apache Eagle, Eagle, Apache Hadoop, Hadoop, Apache HBase, HBase, Apache Hive, Hive, Apache Ambari, Ambari, Apache Spark, Spark, Apache Kafka, Kafka, Apache Storm, Storm, Apache Maven, Maven, Apache Tomcat, Tomcat, Apache Derby, Derby, Apache Cassandra, Cassandra, Apache ZooKeeper, ZooKeeper, Apache, the Apache feather logo, and the Apache project logo are trademarks of The Apache Software Foundation.