Network bandwidth graph check interval

Troubleshooting
1

CMK version:
2.2.0p22
OS version:
Centos Streamline 9

Hi everybody,

i’ve just seen that the performance graphs about the bandwidth consumption have the same check interval of the host (1 minute).
This is very high for some bandwidth monitorings.
I tried to decrease it modifying “Service normal/retry check interval” but no success.

Does someone knows if there a way to have some “Service Perf-O-Meter” checking less than every minute?

Thank you all in advance
Thomas

2

This is because the Interface checks get their data from the agent, either the Checkmk agent or the SNMP agent. An agent query will always get data for all interfaces at once.

You cannot set a different check interval for these checks. You can only set a check interval on the “Check_MK” service check that queries the agent. But then all other checks will only update in this larger interval.

Why is a one minute interval “too high”?

3

Thanks Robert for the detailed reply.

is had a situation where it’d be nicer to have a performance graph updated more times in a minute.
Anyway, i understood what you said.

I understand the implication of setting a lower querying rate, could you tell me how to do it?

Your help is very appreciated
Thomas

4

So you need a higher check interval, not a lower query rate.

This is often not feasible with SNMP as the protocol is very slow.