Audit Slurm queue state to identify resource-stranding job submissions.
npx skills add https://github.com/olliecrow/codex --skill cluster-blame
Please help me install this Agent Skill. Skill: cluster-blame Source: https://github.com/olliecrow/codex/tree/main/skills/cluster-blame Command: npx skills add https://github.com/olliecrow/codex --skill cluster-blame
This Skill helps identify why cluster resources (CPU, memory, GPU) appear idle and who might be blocking scheduling due to misconfigured job submissions.
Use the cluster-blame skill to quickly scan the current Slurm queue and identify likely users or jobs currently stranding CPU, GPU, or memory capacity.
High-intent search queries and answers about installing and using this skill.
Identify Slurm resource stranding by auditing queue state and node packing to pinpoint inefficiently shaped job submissions that block scheduling. This analysis provides evidence-backed attribution distinguishing between user misconfiguration and scheduler policy effects.
Slurm jobs may remain pending with idle GPUs due to resource stranding, where inefficiently shaped job submissions fragment cluster capacity. Auditing node availability and queue state pinpoints specific users whose allocations block scheduling.
Diagnose HPC scheduling bottlenecks by analyzing job shapes, resource allocation, and node availability to provide evidence-backed attribution and confidence scoring for potential blockers. This distinguishes user misconfiguration from scheduler policy effects.
Check Slurm scheduler policy effects by auditing job shapes and node availability. The analysis distinguishes between scheduler policy effects and user misconfiguration, providing confidence-ranked evidence for potential scheduling blockers.
Audit Slurm queue state for capacity planning by analyzing job shapes and resource allocation to identify resource-stranding submissions. This provides evidence-backed attribution showing who is blocking scheduling and why.
Cluster-blame operates without additional dependencies to perform HPC performance analysis. It audits Slurm queue state and node packing internally to identify probable resource-stranding job submissions with confidence scoring.
scitixDiagnose Volcano queue scheduling bottlenecks by analyzing weights, resources, and state.
ovachieverProcess distributed ML data pipelines with Ray Data across CPU/GPU clusters.
QuantumBFSSubmit, monitor, and diagnose Slurm cluster array jobs with pre-submit feasibility checks.
scitixDiagnose resource insufficiency in Volcano-scheduled workloads by analyzing pod requests and node capacity.
