Inbound utilization is very high (> 90 % over 1h) on [...] interface cpu ()
linuxlurak opened this issue · 7 comments
I get this problem popping up all the time:
CPU is measured in kbps (or am i wrong and thats ok?) like this:
I doubt cpu usage is always over 90%... Or do you know a method to check cpu utilization on this switch?
At a glance it looks like one of the ports on your switch is labeled "cpu". I don't think the GST108Tv2 expose CPU utilization over snmp. I'll check to see if I have the same thing on our production.
@linuxlurak Note, that this is not the CPU utilization, but rather the CPU interface. The unit of the values are packets per second.
In case you are using the Template Stack Netgear GS108Tv2 SNMPv2 template, consider to add a discovery filter which only includes the physical and LAG interfaces. Navigate to Configuration > Templates > Template Stack Netgear GS108Tv2 SNMPv2 > Discovery rules > Network interfaces > Filters and add the following filter:
Label Macro Regular expression
A {#SNMPVALUE} matches ^(g|l)[0-9]+$
The above filter should only discover interfaces named g0, g1 g2, [...] and LAGs l0, l1 l2, [...].
HTH.
Thanks for your help. I'll check it out.
@linuxlurak Do you still experience problems with the monitoring of the CPU interface here, or can we close this? Thanks for your feedback.
It works as it seems. "cpu" is not found anymore under "Items" of the device. Do you have a suggestion how to treat existing devices? Just delete the item?
Do you have a suggestion how to treat existing devices? Just delete the item?
Either wait for the specific amount of days configured within the discovery rule under Keep lost resources period (in days) (Configuration / Templates / [TEMPLATE] / Discovery rules / [DISCOVERY RULE]) or temporarily lower it to zero (0) to speed things up.
@linuxlurak can we close this, or are there any issues left? Thanks for your feedback.