Slurm Job Management

Your Job Queue, Now on the Web

All the power of Slurm — without the sbatch, scontrol and sacct sprawl. Submit jobs, watch the queue and report on accounting; all from your browser.

No CLI required

A job submission form, queue filters and one-click cancel/hold/release — without opening a terminal.

Per-user attribution

Every job is submitted under the signed-in user's identity; fairshare and accounting stay accurate.

What's Inside

Full Coverage from Scheduling to Reporting

TULPAR brings Slurm's entire administrator and user surface into one panel.

Job Queue

A live queue view — filter, sort, drill into job details; tell running and pending jobs apart at a glance.

Submission & Templates

Submit jobs from a web form or share parameterized templates with the team — a self-service HPC experience.

Accounting & Analytics

sacct data in the panel: per-user/per-account consumption, job history and analytics charts.

Partitions & QOS

Partition states, QOS definitions, reservations and association management — in one place.

Permissions & Scope

With RBAC, team leads see and manage only their own team's jobs; a fail-closed model.

Service Management

slurmctld, slurmdbd, slurmrestd and munge status; a slurm.conf editor and scontrol reconfigure.

Job Analytics and Deep Visibility

Go beyond the queue: analyze job history, resource consumption and scheduling behavior in the panel.

Jobs are submitted under the signed-in user's POSIX/LDAP identity rather than a fixed service account — fairshare, accounting and permissions stay accurate.

TULPAR talks to Slurm over the slurmrestd REST API and falls back to the CLI when needed — a stable, version-independent interface.

The job detail view joins Slurm accounting data with Prometheus time series — a job's real resource profile in one chart.

Cancel, hold and release jobs from the panel; scontrol subcommands run safely behind an allow-list.

Want to see it live on your cluster? Get in touch!
%