-
Notifications
You must be signed in to change notification settings - Fork 5.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Cannot use instance wildcard in win_perf_counters #2879
Comments
Thanks for the bug report, does it work if you only have "chrome" in the Instances? [[inputs.win_perf_counters]]
[[inputs.win_perf_counters.object]]
ObjectName = "Process"
Instances = ["chrome"]
Counters = [
"% Processor Time"
]
Measurement = "win_proc" |
No it doesn't. Only one metric for the process chrome are reported. |
@bullshit Do you have any idea what might be causing this? |
I used my c++ test code with the \Process(*)\% Processor Time query and it seems that the phd.dll cannot handle the instance names correctly. Thats the output
so telegraf retrieved the instance name correctly from pdh.dll but you don't know if this instance could be e.g. chrome#1, chrome#2 and so on. The same behaviour occurs when you query only chrome instances To avoid this when using |
Just wanted to jumpi in and confirm my original issue #1827 is not solved.
The above still results in very weird percentages for processes that have multiple instances. Single instance services behave normally. |
Same here. Using a wildcard on 'Instances' produces metrics only for the first instance, meanwhile, listening a few instances (es. "chrome","chrome#1","chrome#2"...etc) results in strange behavior, like n-different metrics but not dynamically changing as the new processes come up. I mean, launching telegraph with 2 chrome.exe opened produces 2 metrics but opening a third one does not produce a third metric. I'm not even sure that the connected counter were working correctly. Many thanks! |
Is there any ETA for resolution regarding this issue ? |
@dudusakharovich I don't think anyone is actively working on this, fix will depend on a contribution from the community. |
sure it depends, as pretty much any project here ;) Anyway it looks an important issue to fix, i would love to be able to contribute but.... |
i will take a look at it. But i think it isn't only a quick fix. I have to think about the best solution and that will take some time |
It looks that solving wildcard expanding (#1291) in PR #3018 also resolved this problem. Output on my Windows 10:
When looking at the state of #3031, there are, at least, a several regression problems to solve. I can work at that. |
Bug report
According to #1827 which was released in version 1.3 Windows multi instance processes should be reported as separate metrics, but my tests shows that it doesn't when you gather all instances (*).
Relevant telegraf.conf:
System info:
Telegraf 1.3.1
Windows 10
Steps to reproduce:
Expected behavior:
Actual behavior:
Additional info:
Side note: Using this together with the Prometheus output plugin gives one metric for process chrome for example.
If i change the configuration and specify specific instances it report metrics per process and instance as expected, i.e:
The text was updated successfully, but these errors were encountered: