YARN Services

The YARN Services page displays memory and CPU utilization of YARN processes running on nodes in the cluster over a period of time.

The top panel displays the following metrics.

MetricDescription
Active Resource ManagersThe total number of live resource managers.
Live Node ManagersThe total number of live resource managers.

The following charts in the dashboard contain aggregated values of the metrics. To view usage by cluster, click Show Individuals in the chart of your choice.

Chart NameDescription
YARN RM ProcessThe active state of the YARN resource manager process.
YARN NM ProcessThe active state of the YARN node manager process.
RM Running ThreadsThe number of runnable threads in the resource manager.
NM Running ThreadsThe number of runnable threads in the node manager.
RM Waiting ThreadsThe number of threads waiting in the resource manager. The following states can be tracked:
Threads waiting: The number of threads waiting to execute.
Threads Timed Waiting: The number of threads waiting for other threads to execute for up to a specified waiting time.
NM Waiting ThreadsThe number of threads waiting in the node manager. The following metrics can be tracked:
Threads waiting: The number of threads waiting to execute.
Threads Timed Waiting: The number of threads waiting for other threads to execute for up to a specified waiting time.
CPU Usage Node ManagerThe rate of change of CPU utilization by the NodeManager.
CPU Usage Resource ManagerThe rate of change of CPU utilization by the ResourceManager.
Physical Memory Node ManagerThe amount of physical memory used by the NodeManager process running on a node aggregate shows mean usage by default. Clicking Show Individuals shows actual usage of each process in the node.
Physical Memory Resource ManagerThe amount of physical memory used by Resource Manager process on a node.
Heap Memory Resource ManagerThe amount of heap memory used by the resource manager. The following metrics can be tracked.
Mem Heap Max:The maximum amount of heap memory available to use.
Mem Heap Committed: The amount of heap memory committed to use.Mem Heap Used: The amount of heap memory used.
Heap Memory Node ManagerThe amount of heap memory used by the node memory. The following metrics can be tracked.
Mem Heap Max:The maximum amount of heap memory available to use.
Mem Heap Committed: The amount of heap memory committed to use.Mem Heap Used: The amount of heap memory used.

Resource Allocation Trend

This chart displays the resource trend analysis on hosts showing the following metrics.

  • Allocated Containers: The number of containers allocated to the resources by the resource manager.
  • Allocated Memory: The amount of memory allocated to resources by the resource manager (in GB).
  • Allocated Cores: The number of vCores allocated to resources by the resource manager.
  • Available Memory: The amount of memory available to use.
  • Available Cores: The number of vCores available to use.

Container Summary Trend

This chart displays the summary trend of usage of containers on different hosts showing the following metrics.

  • Containers Completed: The number of containers that completed executing.
  • Containers Running: The number of containers currently running.
  • Containers Failed: The number of containers that failed execution.
  • Containers Killed: The number of containers that were terminated.
  • Containers Launched: The number of containers launched.