Combating Underutilization with the Jobstats Job Monitoring Platform

In 2023, we introduced the Jobstats job monitoring platform which provides user-facing commands and interfaces for inspecting the efficiency of Slurm jobs on CPU and GPU clusters. The platform builds on the Prometheus monitoring framework and the Grafana visualization toolkit. The platform has been adopted by tens of institutions throughout the world. In this poster, we provide updates on the platform, which includes the release of a new component for mitigating underutilization. Job Defense Shield is a software tool for identifying (or even automatically cancelling) user jobs that are underutilizing high-performance computing resources such as GPUs. Users are sent automated email alerts while system administrators can view reports. Job Defense Shield is a tool for both job monitoring and user training.