Identifying Resource-Hungry Workloads

This task compares a workload with all other workloads in your cluster for identifying workloads that are using an excessive amount of resources.

Workload Manager and Workload XM enables you to compare the resource consumption of a particular workload or user against all other workloads or users in the cluster.
For example, you can compare the resource consumption of a specific workload against all the workloads in the cluster with the Resource Consumption chart widget, as shown in the following image:


When you compare the two chart metrics, the chart on the left shows that of a single workload that is consuming half the memory; 1.9 TiB verses 3.7 TiB.

You can delve deeper by creating a Workload View whose user criteria value is the user whose jobs are consuming an inordinate amount of memory and then create a Workload View whose user criteria value comprises all the users except the user that is of interest.


When you compare the two chart metrics over the same time frame, the chart on the left shows the memory consumption as 84 GiB and the CPU core hours as 6.1K compared to 3.7 TiB and 20K.

You could migrate this user to Cloudera and relieve some resources from your on-premises system.