| 1 | # HELP DCGM_FI_DEV_GPU_UTIL GPU utilization (in %). |
| 2 | # TYPE DCGM_FI_DEV_GPU_UTIL gauge |
| 3 | DCGM_FI_DEV_GPU_UTIL{gpu="0",UUID="GPU-aaa",Hostname="host1"} 80 |
| 4 | |
| 5 | # HELP DCGM_FI_DEV_FB_USED Framebuffer memory used (in MiB). |
| 6 | # TYPE DCGM_FI_DEV_FB_USED gauge |
| 7 | DCGM_FI_DEV_FB_USED{gpu="0",UUID="GPU-aaa",Hostname="host1"} 1024 |
| 8 | |
| 9 | # HELP DCGM_FI_DEV_XID_ERRORS Value of the last XID error encountered. |
| 10 | # TYPE DCGM_FI_DEV_XID_ERRORS gauge |
| 11 | DCGM_FI_DEV_XID_ERRORS{gpu="0",UUID="GPU-aaa"} 31 |
| 12 | |
| 13 | # HELP DCGM_FI_DEV_ROW_REMAP_FAILURE Whether remapping of rows has failed. |
| 14 | # TYPE DCGM_FI_DEV_ROW_REMAP_FAILURE gauge |
| 15 | DCGM_FI_DEV_ROW_REMAP_FAILURE{gpu="0",UUID="GPU-aaa"} 1 |
| 16 | |
| 17 | # HELP DCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS Number of remapped rows for uncorrectable errors. |
| 18 | # TYPE DCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS counter |
| 19 | DCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS{gpu="0",UUID="GPU-aaa"} 7 |
| 20 | |
| 21 | # HELP DCGM_FI_DEV_POWER_VIOLATION Throttling duration due to power constraints (in ns). |
| 22 | # TYPE DCGM_FI_DEV_POWER_VIOLATION counter |
| 23 | DCGM_FI_DEV_POWER_VIOLATION{gpu="0",UUID="GPU-aaa"} 2000000 |
| 24 | |
| 25 | # HELP DCGM_FI_DEV_THERMAL_VIOLATION Throttling duration due to thermal constraints (in ns). |
| 26 | # TYPE DCGM_FI_DEV_THERMAL_VIOLATION counter |
| 27 | DCGM_FI_DEV_THERMAL_VIOLATION{gpu="0",UUID="GPU-aaa"} 5000000 |
| 28 | |
| 29 | # HELP DCGM_FI_DEV_MEMORY_TEMP Memory temperature (in C). |
| 30 | # TYPE DCGM_FI_DEV_MEMORY_TEMP gauge |
| 31 | DCGM_FI_DEV_MEMORY_TEMP{gpu="0",UUID="GPU-aaa"} 9223372036854775794 |
| 32 | |
| 33 | # HELP DCGM_FI_PROF_PCIE_TX_BYTES PCIe TX bytes. |
| 34 | # TYPE DCGM_FI_PROF_PCIE_TX_BYTES gauge |
| 35 | DCGM_FI_PROF_PCIE_TX_BYTES{gpu="0",UUID="GPU-aaa"} 123456 |
| 36 | |
| 37 | # HELP DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_TOTAL Total number of NVLink retries. |
| 38 | # TYPE DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_TOTAL counter |
| 39 | DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_TOTAL{gpu="0",gpu_uuid="GPU-aaa",nvlink="1"} 4 |
| 40 | |
| 41 | # MIG sample for the same DCGM_FI_DEV_GPU_UTIL metric family. |
| 42 | DCGM_FI_DEV_GPU_UTIL{gpu="0",UUID="GPU-aaa",GPU_I_ID="2",GPU_I_PROFILE="1g.10gb"} 60 |
| 43 | |
| 44 | # HELP go_gc_duration_seconds A Go runtime metric that should be ignored by the collector. |
| 45 | # TYPE go_gc_duration_seconds summary |
| 46 | go_gc_duration_seconds{quantile="0.5"} 0.001 |