Consumers

Introduction

This plugin groups the jobs based on similar features and displays statistical information about these groups. Available grouping features are:

  • Default features: Account, Application, Comment, GID, Group, JobName, Partition, UID, User, QOS and any other string column kept in the database

  • Advanced features added to jobs using data enhancers during accounting logs ingestion (only features of string type)

How it works

Select a category and a sub-category on which to aggregate the jobs.

The generated table shows for each group:

  • the maximum number of allocated CPUs (Allocated_cores)

  • the maximum execution time (Execution_Time)

  • the total consumed CPU hours (Core_hours)

  • the number of jobs (Count)

  • the average waiting time (Waiting_Time)

  • the average slowdown (Slowdown)

  • the total cost (Cost) in currency as set in the configuration (see Interface configuration)

  • the total consumed energy in kWh (Energy)

  • the total carbon footprint in kg CO₂e (CO₂)

This plugin allows you to get the cluster consumption for each user for example. You can then order the users by clicking on a column name (Core_hours for example) to spot the bigger consumers.

image0

You can click on each group in the table to get more detailed statistics about the above-mentionned variables and its jobs status distribution (see State key metrics for more details).

image1