Skip to main content

Monitoring Submissions

Track job status, view resource metrics, inspect logs, and review Slurm job info.

Monitoring Submissions

The submission detail page tells you everything about a running job: where it is in the pipeline, what resources it's using, and what Slurm thinks of it.

The detail page has five tabs.

Overview tab

Key details displayed:

  • Job Submission ID — copyable
  • Owner — clickable user preview
  • Script — clickable preview
  • Cluster — clickable preview
  • Created — timestamp
  • Last Updated — timestamp
  • Slurm Job ID
  • Slurm Job State
  • Description

Actions: Re-Submit, Delete, Edit.

Abnormal statuses (REJECTED, ABORTED) display a banner with the report message.

Job Metrics tab

Available after approximately 5 data points have been collected.

Charts:

  • CPU utilization, CPU frequency
  • Memory RSS, Memory Virtual
  • GPU utilization, GPU memory
  • Disk read, Disk write
  • Page faults

Controls:

  • Live data toggle — Since start / Last 3 hours / 30 min / 5 min
  • Date range picker
  • Sample rate selector — 10 seconds / 1 minute / 10 minutes / 1 hour
  • Per-node filtering

Data refreshes in real-time for active jobs.

Progress tab

Table of stage transitions through the submission pipeline, from creation to finalization.

Each row shows stage name, status, and timestamp.

Sbatch Arguments tab

Table listing all sbatch arguments passed at submission, including the auto-generated --partition flag.

Slurm Job Info tab

Available after the submission reaches Slurm. Displays a searchable JSON tree viewer showing the raw scontrol show job output. Use it to debug scheduling decisions, resource allocations, or Slurm-level error messages.

Ask AI
Ask a question about Vantage Compute...