Skip to main content


Specifying a cluster ID or name

Throughout the Unravel UI there are pages and tabs where the cluster information is used or needed. For example

  • Various reports, e.g., Chargeback reports.

  • Applications tab where Unravel displays the app's cluster.

When a cluster manager has only a single cluster, Unravel discovers the cluster ID/name automatically. However, when your manager has multiple clusters, you must tell Unravel the clusters' names; allowing you, when necessary, to choose the cluster you want to use.


Use the following port numbers:


Cloudera Manager = 7180
Ambari = 8080
MapR Control System = 8440


Cloudera Manager = 7183
Ambari = 8443
MapR Control System = 8443


Don't specify cluster ID in the javaagent part of sensor properties, i.e., spark.driver.extraJavaOptions and spark.executor.extraJavaOptions

Set in as described below.

  1. Find the cluster's internal "name".

    Example using the Cloudera Manager API using the name field, with fallback to the displayName field.


    Depending on the version of Cloudera, the endpoint may use an API version that is between 16 and 19.

    In Unravel version, the HDFS Capacity Forecasting and Cloud Reports first attempt to get a match on the name attribute, and if no match is found, then search based on the displayName attribute.

         "name": "cluster", 

    Example with Ambari API using the cluster_name field:

          "Clusters": {
            "cluster_name": "DEVHDP26E",

    Example with MapR API using the name field

          "cluster": {
            "name": "QAMAPR60E",
  2. In, set to the exact cluster name you retrieved from the cluster manager. For example:
  3. In, set com.unraveldata.cluster.type to CDH, HDP, or MAPR. If you don't set this, Unravel attempts to determine this automatically if there are artifacts or config files on the Unravel server that may indicate if it contains Ambari (or HDP), Cloudera Manager (or CDH), or MAPR.