vrr/cozystack
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

docs(monitoring): mark gpu-fleet average utilization as legacy NVML

Browse source 
Clarify that the "Average utilization" panel on gpu-fleet reflects the
legacy NVML view (DCGM_FI_DEV_GPU_UTIL) rather than engine-active
profiling. For AI/LLM workloads the NVML number is optimistic; the
gpu-efficiency dashboard carries the profiling-based view.

Signed-off-by: Arsolitt <arsolitt@gmail.com>
This commit is contained in:
Arsolitt 2026-04-18 07:11:28 +03:00
parent 2518e09d67
commit 5b210ac7fd
No known key found for this signature in database
GPG key ID: 4D8302CE6A9247C4
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
dashboards/gpu/gpu-fleet.json
View file

@ -268,7 +268,7 @@
}
],
"title": "Average utilization",
"description": "NVML utilization across all tenant GPUs (engine-active %).",
"description": "Legacy NVML utilization across all tenant GPUs.",
"transparent": false,
"datasource": {
"type": "prometheus",