CPU percentage usage- odd values and how to discard those

[majster]

Hi,
I am monitoring container applications. and one of them has some odd values of CPU Percent Usage.
There is an lets say 20 good readings, and one that is E+12 higher than the others. It makes my graphs completely unreadable.

Look at the sample of Last values from it.

2023-05-17 07:41:50 0
2023-05-17 07:40:50 1186285791323.91
2023-05-17 07:39:50 188.9595
(...)
2023-05-17 08:22:50 0.00006464
2023-05-17 08:21:50 1254880548002.09
2023-05-17 08:20:50 119.7664
(...)
2023-05-17 07:10:50 177.3935
2023-05-17 07:09:51 997121300342.394
2023-05-17 07:08:50 191.781

I don't know how are those values come so big, but I am wondering, am I able to not to take those under consideration when creating a Graph? Like "I don't care about values higher than 1000" f.e. ?

[tim.mooney]

You don't say what version of Zabbix you're using, which may be important for how to answer the question.

Assuming you are using a version of Zabbix new enough to support "item value preprocessing", look at the "Validation" section of the docs for preprocessing: https://www.zabbix.com/documentation.../preprocessing

[Hamardaban]

Most likely "preprocessing" >> "Validation" >> "In range" ..