aarch64 麒麟v10系统使用docker部署nvidia_gpu_exporter监控GPU

nvidia_gpu_exporter

 随着科技的发展，人工智能对于图像的处理越来越频繁，对于图像处理的场景也越来越多，GPU的使用也随之而来；那对于性能测试的图像比对、3D处理等类型的系统使用到了GPU，我们需要对GPU进行细化了解和监控。

 nvidia_gpu_exporter是prometheus 的 Nvidia GPU 导出器，使用nvidia-smi二进制文件收集指标。

容器启动命令

docker run -d \
--name nvidia_smi_exporter \
--restart unless-stopped \
--device /dev/nvidiactl:/dev/nvidiactl \
--device /dev/nvidia0:/dev/nvidia0 \
-v /usr/lib64/libnvidia-ml.so:/usr/lib/aarch64-linux-gnu/libnvidia-ml.so \
-v /usr/lib64/libnvidia-ml.so.1:/usr/lib/aarch64-linux-gnu/libnvidia-ml.so.1 \
-v /usr/bin/nvidia-smi:/usr/bin/nvidia-smi \
-p 9835:9835 \
utkuozdemir/nvidia_gpu_exporter:1.1.0

 github

GitHub - utkuozdemir/nvidia_gpu_exporter: Nvidia GPU exporter for prometheus using nvidia-smi binary

grafana展示

，面板链接Nvidia GPU Metrics | Grafana Labs

 

 采集数据

# TYPE nvidia_smi_clocks_throttle_reasons_sync_boost gauge
nvidia_smi_clocks_throttle_reasons_sync_boost{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_command_exit_code Exit code of the last scrape command
# TYPE nvidia_smi_command_exit_code gauge
nvidia_smi_command_exit_code 0
# HELP nvidia_smi_compute_mode compute_mode
# TYPE nvidia_smi_compute_mode gauge
nvidia_smi_compute_mode{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_count count
# TYPE nvidia_smi_count gauge
nvidia_smi_count{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 1
# HELP nvidia_smi_display_active display_active
# TYPE nvidia_smi_display_active gauge
nvidia_smi_display_active{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_display_mode display_mode
# TYPE nvidia_smi_display_mode gauge
nvidia_smi_display_mode{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 1
# HELP nvidia_smi_ecc_errors_corrected_aggregate_device_memory ecc.errors.corrected.aggregate.device_memory
# TYPE nvidia_smi_ecc_errors_corrected_aggregate_device_memory gauge
nvidia_smi_ecc_errors_corrected_aggregate_device_memory{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_aggregate_dram ecc.errors.corrected.aggregate.dram
# TYPE nvidia_smi_ecc_errors_corrected_aggregate_dram gauge
nvidia_smi_ecc_errors_corrected_aggregate_dram{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_aggregate_sram ecc.errors.corrected.aggregate.sram
# TYPE nvidia_smi_ecc_errors_corrected_aggregate_sram gauge
nvidia_smi_ecc_errors_corrected_aggregate_sram{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_aggregate_total ecc.errors.corrected.aggregate.total
# TYPE nvidia_smi_ecc_errors_corrected_aggregate_total gauge
nvidia_smi_ecc_errors_corrected_aggregate_total{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_volatile_device_memory ecc.errors.corrected.volatile.device_memory
# TYPE nvidia_smi_ecc_errors_corrected_volatile_device_memory gauge
nvidia_smi_ecc_errors_corrected_volatile_device_memory{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_volatile_dram ecc.errors.corrected.volatile.dram
# TYPE nvidia_smi_ecc_errors_corrected_volatile_dram gauge
nvidia_smi_ecc_errors_corrected_volatile_dram{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_volatile_sram ecc.errors.corrected.volatile.sram
# TYPE nvidia_smi_ecc_errors_corrected_volatile_sram gauge
nvidia_smi_ecc_errors_corrected_volatile_sram{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0
# HELP nvidia_smi_ecc_errors_corrected_volatile_total ecc.errors.corrected.volatile.total
# TYPE nvidia_smi_ecc_errors_corrected_volatile_total gauge
nvidia_smi_ecc_errors_corrected_volatile_total{uuid="cb792315-81df-27cd-cc86-78f7ff647b43"} 0

​​​​​​​