Hello
First, sorry for my english.
I have created a custom cpu poller for linux server because the native one just show the load of CPU cores individually, and i want the real load average.
But I have a problem : when the server is in full load, the snmp collect for CPU & RAM metrics doesn't work.
example :
With the standard CPU poller, whenever the CPU is fully charged, the CPU metrics are correctly collected.
But with my custom poller, i have graph holes when the server is in high load :
(CPU # 1= load1, CPU # 2= load5, CPU # 3= load15)
My custom CPU & Memory poller :
(i count the number of CPU cores : with a 16 cpu cores, if the load average = 16 >>>> 100% ; with a 2 cpu cores, if the load average = 1 >>>> 50%, etc)
Can you explain me why with my custom CPU poll i have graph hole, but with standard CPU poll not ?
The aim is to graph load15 according to the number of cpu cores. If you have a better solution, tell me (1.3.6.1.4.1.2021.10.1.4 is not used).
thanks