mirror of
https://github.com/Alkatrazz24/Funk-lab.git
synced 2026-07-08 11:04:43 +02:00
feat(monitoring): métriques conso électrique CPU+GPU (RAPL + ROCm power)
- node_exporter : activer --collector.rapl (CPU power package, storage-01 + gpu-01 + nœuds k8s) - rocm_scraper : ajouter rocm_gpu_power_watts + rocm_gpu_power_cap_watts via sysfs hwmon - gpu-01 : activer rocm_exporter_enabled - kube-prometheus-stack : propager --collector.rapl aux DaemonSet Talos via prometheus-node-exporter.extraArgs Déployé sur storage-01 et gpu-01. ArgoCD prend en charge les nœuds Talos. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
parent
8236448859
commit
91e648f55e
4 changed files with 19 additions and 1 deletions
|
|
@ -9,7 +9,8 @@ ExecStart=/usr/local/bin/node_exporter \
|
|||
--web.listen-address=":{{ node_exporter_port }}" \
|
||||
--collector.textfile.directory={{ node_exporter_textfile_dir }} \
|
||||
--collector.systemd \
|
||||
--collector.processes
|
||||
--collector.processes \
|
||||
--collector.rapl
|
||||
|
||||
Restart=on-failure
|
||||
RestartSec=5s
|
||||
|
|
|
|||
|
|
@ -12,6 +12,10 @@ TMPFILE="${OUTPUT}.tmp"
|
|||
echo "# TYPE rocm_vram_used_bytes gauge"
|
||||
echo "# HELP rocm_vram_total_bytes VRAM total in bytes"
|
||||
echo "# TYPE rocm_vram_total_bytes gauge"
|
||||
echo "# HELP rocm_gpu_power_watts GPU power draw in watts"
|
||||
echo "# TYPE rocm_gpu_power_watts gauge"
|
||||
echo "# HELP rocm_gpu_power_cap_watts GPU power cap in watts"
|
||||
echo "# TYPE rocm_gpu_power_cap_watts gauge"
|
||||
|
||||
for card in /sys/class/drm/card[0-9]; do
|
||||
[ -d "$card/device" ] || continue
|
||||
|
|
@ -29,5 +33,11 @@ TMPFILE="${OUTPUT}.tmp"
|
|||
|
||||
vram_total=$(cat "$card/device/mem_info_vram_total" 2>/dev/null)
|
||||
[ -n "$vram_total" ] && echo "rocm_vram_total_bytes{${labels}} ${vram_total}"
|
||||
|
||||
power_raw=$(cat "$card/device/hwmon/hwmon"*/power1_average 2>/dev/null | head -1)
|
||||
[ -n "$power_raw" ] && echo "rocm_gpu_power_watts{${labels}} $(echo "scale=2; ${power_raw}/1000000" | bc)"
|
||||
|
||||
power_cap_raw=$(cat "$card/device/hwmon/hwmon"*/power1_cap 2>/dev/null | head -1)
|
||||
[ -n "$power_cap_raw" ] && echo "rocm_gpu_power_cap_watts{${labels}} $(echo "scale=2; ${power_cap_raw}/1000000" | bc)"
|
||||
done
|
||||
} > "$TMPFILE" && mv "$TMPFILE" "$OUTPUT"
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue