You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Telegraf version: custom build based on 1.2.1 with some master & PR commits cherry-picked. phemmer@1c214d0
OS: Linux, CentOS/7
Steps to reproduce:
Run telegraf with procstat input
Have a monitored process use no cpu between polling interval.
Expected behavior:
Report a 0 value for cpu_usage.
Actual behavior:
Reports a null value for cpu_usage.
Additional info:
A 0 value cpu usage is not the same as a null value cpu usage. Null means we have no data. 0 means we have data, but it was 0. If you want to do something like calculate an average, you'll get incorrect results as it wont factor in the points where CPU usage was really 0.
I personally have no intention of using cpu_usage, as I much prefer to use the raw time metrics and calculate CPU usage myself. Just reporting as it is a bug as it will affect anyone using it.
The text was updated successfully, but these errors were encountered:
Bug report
Relevant telegraf.conf:
System info:
Telegraf version: custom build based on 1.2.1 with some master & PR commits cherry-picked. phemmer@1c214d0
OS: Linux, CentOS/7
Steps to reproduce:
Expected behavior:
Report a
0
value forcpu_usage
.Actual behavior:
Reports a null value for
cpu_usage
.Additional info:
A
0
value cpu usage is not the same as a null value cpu usage. Null means we have no data.0
means we have data, but it was 0. If you want to do something like calculate an average, you'll get incorrect results as it wont factor in the points where CPU usage was really 0.I personally have no intention of using
cpu_usage
, as I much prefer to use the raw time metrics and calculate CPU usage myself. Just reporting as it is a bug as it will affect anyone using it.The text was updated successfully, but these errors were encountered: