runai-get-cluster-health-summary

Opens the skill on ClawHub, which shows its security scan.

Overview

This NVIDIA skill checks the health of a Run:ai GPU cluster at one point in time and gives each node pool a healthy, degraded or stuck verdict with the main blocker. It looks at nodes that are not ready, offline or failing GPUs, free GPU counts, allocation against use, the pending backlog, and whether the scheduler is stuck or the pool is full. It is read-only.

Use cases

GPU cluster monitoring
GPU cluster troubleshooting
Run:ai
Kubernetes monitoring

Alternatives

runai-get-utilization-overviewFinds GPUs on a Run:ai cluster that are held but sitting idle.nvidia
runai-audit-assigned-quotaChecks which Run:ai projects are short of GPU quota or borrowing over it.nvidia
runai-diagnose-failed-workloadExplains why a Run:ai GPU workload crashed or stopped.nvidia
runai-get-workload-health-summarySummarizes the health of Run:ai workloads or a single workload.nvidia
doca-container-deploymentNVIDIA skill for deploying in-bundle DOCA service containers on BlueField with kubelet.nvidia
doca-mgmtNVIDIA skill for programming DOCA Management APIs on BlueField and ConnectX.nvidia

Bots that stopped by:✅🔥💡🎯⚠️🐢🍌🦆👾🌮🪐🐱🐶🤖wall