sendnrw/windows_exporter
2
0
Fork 0
mirror of https://github.com/prometheus-community/windows_exporter.git synced 2026-10-17 18:19:04 +02:00
Code Issues Packages Projects Releases Wiki Activity

collector: publish base units and _total counter names (#2743)

Browse Source 
Signed-off-by: Jan-Otto Kröpke <mail@jkroepke.de>
This commit is contained in:
Jan-Otto Kröpke
2026-10-16 10:48:26 +02:00
committed by GitHub
parent dc8092f46d
commit 3fb18521f6
131 changed files with 1674 additions and 1459 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/collector.gpu.md
+7 -7
View File
@@ -61,13 +61,13 @@ logged at debug level and never fail the scrape.
| `windows_gpu_engine_frequency_max_hertz` | Maximum clock frequency of the GPU engine in hertz, while not overclocked | gauge | `device_id`,`luid`,`phys`,`eng` |
The engine frequency metrics are only exposed for engines that report a maximum frequency, typically the 3D engine. The
`eng` label matches the `eng` label of `windows_gpu_engine_time_seconds`.
`eng` label matches the `eng` label of `windows_gpu_engine_time_seconds_total`.
### Per-process Metrics
| Name | Description | Type | Labels |
|----------------------------------------------|-------------------------------------------------|---------|-----------------------------------------------------------|
| `windows_gpu_engine_time_seconds` | Total running time of the GPU engine in seconds | counter | `device_id`,`luid`,`phys`, `eng`, `engtype`, `process_id` |
| `windows_gpu_engine_time_seconds_total` | Total running time of the GPU engine in seconds | counter | `device_id`,`luid`,`phys`, `eng`, `engtype`, `process_id` |
| `windows_gpu_process_memory_committed_bytes` | Total committed GPU memory in bytes per process | gauge | `device_id`,`luid`,`phys`,`process_id` |
| `windows_gpu_process_memory_dedicated_bytes` | Dedicated GPU memory usage in bytes per process | gauge | `device_id`,`luid`,`phys`,`process_id` |
| `windows_gpu_process_memory_local_bytes` | Local GPU memory usage in bytes per process | gauge | `device_id`,`luid`,`phys`,`process_id` |
@@ -118,7 +118,7 @@ windows_gpu_adapter_memory_dedicated_bytes{phys="0"}
```promql
sum by (phys) (
rate(windows_gpu_engine_time_seconds{phys="0", engtype="3D"}[1m])
rate(windows_gpu_engine_time_seconds_total{phys="0", engtype="3D"}[1m])
) * 100
```
@@ -126,7 +126,7 @@ sum by (phys) (
```promql
sum by (phys, process_id) (
rate(windows_gpu_engine_time_seconds{process_id="1234", engtype="3D"}[1m])
rate(windows_gpu_engine_time_seconds_total{process_id="1234", engtype="3D"}[1m])
) * 100
```
@@ -144,7 +144,7 @@ sum by (luid, process_id) (
```promql
topk(5, sum by (process_id) (
rate(windows_gpu_engine_time_seconds[1m])
rate(windows_gpu_engine_time_seconds_total[1m])
) * 100)
```
@@ -159,7 +159,7 @@ sum by (phys) (
Show GPU engine time with process owner and command line:
```promql
windows_gpu_engine_time_seconds * on(process_id) group_left(owner, cmdline) windows_process_info
windows_gpu_engine_time_seconds_total * on(process_id) group_left(owner, cmdline) windows_process_info
```
**Show the engine clock as a ratio of its maximum clock:**
@@ -177,7 +177,7 @@ windows_gpu_engine_frequency_hertz / windows_gpu_engine_frequency_max_hertz
- alert: HighGpuUtilization
expr: |
sum by (process_id) (
rate(windows_gpu_engine_time_seconds[1m])
rate(windows_gpu_engine_time_seconds_total[1m])
) * 100 > 80
for: 10m
labels: