Add support for SR-IOV GPU stats (#15796)

* Add option to treat GPU as SRIOV in order for stats to work correctly

* Add to intel docs

* fix tests
This commit is contained in:
Nicolas Mowen
2025-01-03 09:43:44 -06:00
parent d5fb957503
commit 39ecffbf92
6 changed files with 25 additions and 7 deletions

View File

@@ -175,6 +175,16 @@ For more information on the various values across different distributions, see h
Depending on your OS and kernel configuration, you may need to change the `/proc/sys/kernel/perf_event_paranoid` kernel tunable. You can test the change by running `sudo sh -c 'echo 2 >/proc/sys/kernel/perf_event_paranoid'` which will persist until a reboot. Make it permanent by running `sudo sh -c 'echo kernel.perf_event_paranoid=2 >> /etc/sysctl.d/local.conf'`
#### Stats for SR-IOV devices
When using virtualized GPUs via SR-IOV, additional args are needed for GPU stats to function. This can be enabled with the following config:
```yaml
telemetry:
stats:
sriov: True
```
## AMD/ATI GPUs (Radeon HD 2000 and newer GPUs) via libva-mesa-driver
VAAPI supports automatic profile selection so it will work automatically with both H.264 and H.265 streams.

View File

@@ -815,11 +815,13 @@ telemetry:
- lo
# Optional: Configure system stats
stats:
# Enable AMD GPU stats (default: shown below)
# Optional: Enable AMD GPU stats (default: shown below)
amd_gpu_stats: True
# Enable Intel GPU stats (default: shown below)
# Optional: Enable Intel GPU stats (default: shown below)
intel_gpu_stats: True
# Enable network bandwidth stats monitoring for camera ffmpeg processes, go2rtc, and object detectors. (default: shown below)
# Optional: Treat GPU as SR-IOV to fix GPU stats (default: shown below)
sriov: False
# Optional: Enable network bandwidth stats monitoring for camera ffmpeg processes, go2rtc, and object detectors. (default: shown below)
# NOTE: The container must either be privileged or have cap_net_admin, cap_net_raw capabilities enabled.
network_bandwidth: False
# Optional: Enable the latest version outbound check (default: shown below)