Reports for System Administrators

The general idea to generate a report is to simply add the --report option to an alert. Here is an example:

$ job_defense_shield --zero-gpu-util --days=7 --report

With the appropriate entry in config.yaml:

zero-gpu-utilization:
  file: alert/zero_util_gpu_hours.py
  clusters:
    - della
  partitions:
    - gpu
  excluded_users:
    - aturing
    - einstein