Usage

Local Monitoring

Monitor your local machine’s resources in real-time.

# Start interactive monitoring with default 2-second refresh
jobscope

# Custom refresh period (in seconds)
jobscope --period 1.0

# Run once and exit (no continuous monitoring)
jobscope --once

# Write a JSON summary on exit
jobscope --summary ./metrics-summary.json

Press q to quit and clean up. Press Enter or Esc to toggle between global and per-node views.

Slurm Monitoring

Attach to a running Slurm job by ID.

# Monitor a specific Slurm job
jobscope --jobid 123456

# Monitor with custom period and write a JSON summary on exit
jobscope --jobid 123456 --period 5.0 --summary ./job_123456_summary.json

JobScope will wait for pending jobs to start, then run the monitoring worker on the allocated compute nodes using srun while streaming metrics back to your terminal.

Press q to quit and clean up. Press Enter or Esc to toggle between global and per-node views.