NVIDIA/DCGM

Issues somebody may already be working on are listed last. List everything by date

• NVLink health watch reports DCGM_FR_NVLINK_ERROR_CRITICAL for successful link recoveries, and the incident does not clear.

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 75/100

  NVIDIA/DCGM#318 · Opened Oct 8, 2026

• DCGM 4,7.0 reporting Warning when the max power usage is 134 on L40S

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 45/100

  NVIDIA/DCGM#317 · Opened Oct 7, 2026

• After a failed `lshw` parse in `SysMon`, the CPU hierarchy call never returns, so `dcgm-exporter` never binds its port

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 52/100

  NVIDIA/DCGM#315 · 1 comment · 1 reaction · Opened Sep 22, 2026

• DCGM 4.7.0 dcgmi health returns "API version mismatch" on some hosts

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/DCGM#314 · 2 comments · Opened Sep 21, 2026

• DCGM API calls block past 45s on a GPU generating correctable ECC errors under load, and return in 0s when it is idle

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 38/100

  NVIDIA/DCGM#313 · 7 comments · Opened Sep 7, 2026

• RTX Pro 5000 Blackwell mislabelled in diag-skus.yaml.in

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 76/100

  NVIDIA/DCGM#312 · Opened Sep 1, 2026

• Can DCGM Profiling Metrics retrieve metrics at a frequency of less than 100 ms?

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 42/100

  NVIDIA/DCGM#310 · 1 comment · Opened Aug 26, 2026

• `DCGM_FR_IMEX_UNHEALTHY` false positives set on DRA enabled clusters

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#308 · Opened Aug 13, 2026

• [Feature Request] Ship DCGM packages in the NVIDIA driver local repository installer

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/DCGM#307 · 4 reactions · Opened Jul 23, 2026

• dcgmi stats output incorrectly labels "GPU Engine utilisation" as "SM Utilisation"

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 55/100

  NVIDIA/DCGM#305 · Opened Jul 6, 2026

• `dcgmi diag` should run nccl-tests on multiple GPUs

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 68/100

  NVIDIA/DCGM#304 · Opened Jun 24, 2026

• DCGM profiling metrics report 0 after Xid 31 while GPU remains active; nv-hostengine GetSamples -37

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 45/100

  NVIDIA/DCGM#303 · 1 comment · 1 reaction · Opened Jun 17, 2026

• dcgm 4.4.2 test name changed?

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 35/100

  NVIDIA/DCGM#296 · Opened May 5, 2026

• dcgmi InstallCtrlHandler unhandled return

  May be free again A pull request for this issue was closed without being merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 72/100

  NVIDIA/DCGM#294 · Opened May 1, 2026

• Please cut tags and GitHub releases for new versions

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#293 · 4 reactions · Opened May 1, 2026

• Update the repo with 4.5.2 changes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 38/100

  NVIDIA/DCGM#292 · 2 comments · Opened Apr 30, 2026

• How to correctly use the nv-hostengine on Kubernetes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  NVIDIA/DCGM#287 · Opened Mar 24, 2026

• MonitorPower health check produces false positive warnings due to zero-threshold violation time check

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/DCGM#285 · 2 reactions · Opened Mar 8, 2026

• Memory diagnostic may detect a GPU memory allocation error but still mark the test as PASS

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 45/100

  NVIDIA/DCGM#284 · Opened Mar 6, 2026

• Misleading field label "GPU Device IDs Detected" in dcgmi diag output

  May be free again A pull request for this issue was closed without being merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/DCGM#282 · Opened Mar 4, 2026

• Hang detection feature introduced in 4.4.2 floods logs

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/DCGM#279 · 1 comment · Opened Feb 10, 2026

• Can we use this on one FIPS enabled instance?

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  NVIDIA/DCGM#278 · Opened Feb 6, 2026

• Xid Errors aren't reset back unless the driver is reset or node rebooted

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 45/100

  NVIDIA/DCGM#277 · Opened Feb 3, 2026

• bug: persistent/leaked file handles to MIG devices prevents MIG destroys

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#276 · Opened Jan 28, 2026

• Authentication when connecting to remote host

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  NVIDIA/DCGM#275 · Opened Jan 25, 2026

• mla tool fails to parse dcgm_eud.mle with "Failed to find Mle header"

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#274 · Opened Jan 22, 2026

• DCGM 4.5.0: nvbandwidth fails in runlevel 3/4 with "MCUTIL > 10%" immediately after Diagnostic (GpuBurn) even on idle systems; nvbandwidth passes when run standalone

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/DCGM#272 · 3 comments · Opened Jan 19, 2026

• dcgmi health check error,but temp is ok

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#270 · 1 comment · 1 reaction · Opened Dec 22, 2025

• Memory usage doubled compared to v3

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#269 · 1 reaction · Opened Dec 17, 2025

• No GPU Max Memory Bytes reported on L4 GPU

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#268 · Opened Dec 12, 2025

• [bug] The metric DCGM_FI_DEV_ECC_SBE_VOL_TOTAL cannot be collected alongside metrics of type DCGM_FI_DEV_NVLINK_COUNT_*

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#265 · Opened Nov 10, 2025

• Validate NVML_FI_DEV_GET_GPU_RECOVERY_ACTION gpu state during health check

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 45/100

  NVIDIA/DCGM#263 · Opened Nov 6, 2025

• Bug in DCGM Python3 bindings

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 38/100

  NVIDIA/DCGM#262 · Opened Oct 30, 2025

• dcgmbuild/build.sh error

  May be free again A pull request for this issue was closed without being merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 48/100

  NVIDIA/DCGM#260 · 2 comments · Opened Oct 22, 2025

• Can DCGM independently view the health information of the CPU？

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  NVIDIA/DCGM#258 · Opened Oct 9, 2025

• Can dcgm-exporter collect the POD gpu usage metrics in K8S?

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#256 · 1 comment · Opened Sep 20, 2025

• No datacenter-gpu-manager-4-cuda13 Available for Opensuse15

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#255 · Opened Sep 19, 2025

• pcie diagnostic fails on driver 580 with CUDA 12 support packages

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#254 · 1 comment · 2 reactions · Opened Sep 8, 2025

• diag warning after upgrade to 4.2.3

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#253 · 1 comment · 1 reaction · Opened Aug 25, 2025

• Tensor Pipe metrics are inconsistent and have holes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/DCGM#252 · 1 comment · Opened Aug 22, 2025

• missing FP8 & FP4/6 Tensor Active metrics

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 45/100

  NVIDIA/DCGM#251 · 4 comments · Opened Aug 19, 2025

• Is CUDA 11 required for compilation and installation? I am using CUDA 12.8.

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/DCGM#250 · 1 comment · Opened Aug 18, 2025

• dcgmi dmon doesn't work properly with different MIG instances

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#248 · 2 comments · Opened Aug 5, 2025

• Running dcgm diagnostics with PCIe/Nvlink test keep hanging

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#247 · Opened Jul 31, 2025

• dcgm >= 4.0 installs headers in possible wrong include path

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 35/100

  NVIDIA/DCGM#246 · 1 comment · Opened Jul 16, 2025

• Outdated Ubuntu install instructions in README.md

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 1/5 Under an hour Newbie friendliness 50/100

  NVIDIA/DCGM#245 · Opened Jul 15, 2025

• SM active cannot be collected when sampled alone, but becomes available when collected alongside other profiling metrics

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#244 · Opened Jul 15, 2025

• [Feature Request] Export RANK env variable as Prometheus metric in dcgm-exporter

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 35/100

  NVIDIA/DCGM#243 · Opened Jul 14, 2025

• DCGM GPU Diagnostics tests thermal violation too sensitve

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 35/100

  NVIDIA/DCGM#242 · Opened Jul 7, 2025

• [Feature Request] Add Process Information Fields to DCGM Exporter for GPU Allocation Detection

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/DCGM#241 · 1 comment · 1 reaction · Opened Jul 6, 2025