Applications Explorer

The Applications Explorer board displays the status of applications that run on YARN. You can do the following using this dashboard.

  • View inconsistent applications or incidents.
  • View recommendations for faulty apps.

Filtering

In Applications Explorer, click any filter and choose a criteria to display the results. You can filter applications by the following criteria.

FilterDescription
QueuesThe list of queues used in YARN.
App TypeThe type of application submitted to YARN for execution.
StatusThe status of the job.
UsersThe user executing the job in the application.

Searching a Record

To search a record in Applications Explorer, do the following.

  1. Click the search bar located in the top of the dashboard.
  2. Type the name of the field along with the value you are searching for.
    For example, if you are searching for application ID application_1595970633553_0547, type id:application_1595970633553_0547 in the search bar and press enter.

Recommendations

For jobs of every application type you submit, you can view recommendations to optimize the performance of your jobs. These recommendations are based on jobs submitted by applications in the YARN queue of your Hadoop cluster.

Viewing Recommendations

To view recommendations for optimization of jobs, do the following.

  1. Click Recommendations button in the top.

A Recommendations tile is displayed. The recommendation is displayed in the following format.
Number of applications, potential problem category.

For example, 2 Apps, High memory status means that two applications might have high memory issues.

  1. In the Recommendations tile, click the recommendations for the problem category you want to view.

The IDs of applications that have recommendations in that category are displayed.

  1. Click the application ID in the table to view the recommendation.

Viewing Incidents

To view incidents in the application explorer, do the following.

  1. Click Incidents button in the top.

An Incidents tile is displayed. The incident is displayed in the following format.
Number of incidents, service name whose incident was reported, **View incidents** link

For example, 2 Incidents, Hadoop Service, View Incidents status means that two incidents were raised by Hadoop service.

  1. Click View incidents to know more.

The Incidents window is displayed.

Incidents

The Incidents page displays the list of incidents raised by a service. Each incident is displayed in the following format.

(Incident name ) (Time start time and end time of the incident) (The host on which the incident was raised) (Severity of the incident) (Status of the incident, whether cleared or raised)

The following image is an example of the incident format.

Incident Format

Viewing Incident Details

To view incident details, click the incident name to expand the details.

A Details and Alert tab are displayed.

  • The Details tab displays all details of the incident.
  • The Alert tab lets you edit the alert raised by the incident. Click the Edit button to edit the fields of the alert.
  • From the incident details window you can also do the following to view the date and time at which the incident was raised and evaluated.
    1. Click Raised at to expand detailed information of the date and time at which the incident raised.
    2. Click Evaluated at to expand detailed information of the incident evaluation.

Metrics

By default, the table in the dashboard lets view the following metrics of jobs carried out by applications installed in YARN.

MetricDescription
IDThe ID of the application.
VarianceThe value that displays by how many times the application is running slow.
TypeThe type of application.
Final StatusThe status of the application when the associated job completed execution. The final status can be either of the following.
Succeeded, Failed, Killed, Undefined, Unknown.
NameThe name of the job.
UserThe name of the user executing the job.
MemoryThe amount of memory consumed by the job.
vCoresThe amount of vCores used by the application.
DurationThe time consumed to complete the job.
QueueThe name of the queue the application resides in.
StartedThe time at which the job in the application started.
EndedThe time at which the job in the application ended.
cpuUsePercentThe amount of CPU used (in %) by the application.
avgCoreThe average number of cores used.
avgMemoryMbThe average amount of memory used (in MB).
maxMemoryMbThe maximum amount of memory used (in MB).
memoryUsePercentThe amount of memory used (in %).
containersThe number of containers used by the application.
hostsThe list of hosts the application executed jobs on.
memoryMbSecThe rate of memory utilization.
coresSecThe rate of core utilization.
totalTimeThe total time taken to execute the job.

The metrics table might display the following icons for a record. Click these icons to know more.

  • The alert icon displays the service(s) of the application that might require attention.
  • Displays more details of the selected application ID.
  • Displays the following options.

Add to Compare: Compares any two jobs to list the configuration difference. Choose this option for any two jobs in the table that you want to compare.
View Concurrency: Displays the list of jobs that were running at the same time as the selected application ID.

Viewing Logs

For every record, you can view the status log. To view the logs, click the icon in the Status column of the record whose log you want to view, as shown in the following screenshot.

Log

Adding/Removing Columns

To add a column to the Applications Explorer table or to remove a column, do the following.

  1. In the top right column of the table, click the Table Settings icon .
    A Select Column window appears.

  2. From the available list of columns, select the column name you want to add to the table or deselect a column name you want to remove from the table.

  3. Close the Select Column window.
    The table is modified.