Luxe Vantage

A usage-management and billing cloud built on AI/HPC job schedulers. It analyzes resource usage by user, group, and project on Slurm and Kubernetes, and automates status reporting and billing.

Request a LIVE DEMO Get the white paper

LuxeVantage/usageLIVE DEMO · L3

Luxe Vantage usage explorer — Slurm usage filtered by period and grouped by account, with CPU/GPU hours, memory, job counts and chargeback cost per account in USD

↑ Luxe Vantage live operations console (operator view)

01PAIN POINTS

Solving these challenges

The operational hurdles teams hit most often in the field — see how Luxe Vantage resolves them in the LIVE DEMO.

01

Ten departments share the same GPU pool, but the basis for splitting costs is unclear

02

You want to bill by usage like a cloud, but there’s no suitable tool

03

Commercial tools like Run:ai carry heavy cloud dependencies and integrate poorly with Slurm environments

02FEATURES

Feature details

Key capabilities organized by category. Each group is delivered as a working scenario in the LIVE DEMO.

F.01

Unified Scheduler Metrics

Combine Slurm sacct with Kubernetes metrics into a single workload fact table

Slurm — real-time collection of sacct, sreport, and sshare data
Kubernetes — unified Volcano, GPU Operator, cAdvisor, and Run:ai metrics
Automatically maps time-sliced usage of the same GPU pool
Analyze by job (job_id), user, group, project, and queue dimensions

F.02

Usage Analytics

Multi-axis usage analysis by user, group, project, time window, GPU model, and more

Track GPU, CPU, and memory usage by hour, day, week, and month
Analyze the distribution of job wait, run, and fail times (queue · run · fail latency)
Automatically detect inefficient jobs — low GPU utilization, long-running idle, and more
Cross-tab drill-down (user × GPU model × queue)

F.03

Billing & Chargeback Automation

Apply priority weighting and discounts to hourly rates and issue invoice PDFs automatically

Rate-catalog management — register hourly rates by GPU model and queue priority
Priority weighting — differentiated tiers such as premium, standard, and best-effort
Annual, quarterly, and monthly commit discounts with true-up settlement
Automatic KRW · USD exchange-rate application and tax handling
Per-department invoice PDFs — automatic mapping of PO numbers and cost codes

F.04

Budget, Policy & Multi-tenancy

Quarterly budgets, resource quotas, and automatic queue limits head off inter-department resource disputes

Set quarterly budgets with automatic alerts at 80% and 100% thresholds
Automatic queue limits when over budget (integrated with Slurm Account and Kubernetes ResourceQuota)
FairShare policy — automatically adjusts priority based on historical usage
RBAC and data isolation — option to keep usage private between departments

03SCENARIO

Flagship scenario · integrations

Flagship scenario

Resource disputes across 10 departments → automated settlement

−90%

End-of-month manual settlement cut by 90%. Department-head disputes end with "look at your invoice."

INTEGRATIONS

Systems you can integrate

Slurm sacct
Kubernetes (Volcano / GPU Operator)
Luxe Vision

Next steps

Need infrastructure design consulting to make this product possible?

Request a LIVE DEMO View related Solutions