Trouble Monitoring ifHCOutOctets and ifHCInOctets...

[poskalp]

Hi, I'm having some troubles with Zabbix monitoring IF-MIB::ifHCOutOctets and IF-MIB::ifHCInOctets. The graph with the trouble is in the attachment, like you can see every half and hour zabbix graph some lines that are weird and not true, we compare with other snmp monitoring system.

The trouble begin when we start to monitoring the counter64 HC OIDs. In the attachment is the configuration of the items of the graphs, so, i have an error in my configuration? when we monitoring counter64 HC values we have to change something?

I check the server and there is more than enough cpu and ram free.

Thanks in advance for the help.

Greetings

[gleepwurp]

Hi poskalp,

SNMP queries every 20 seconds seems a bit aggressive if you're monitoring a telecom device with potentially a lot of interfaces.

First thing I'd try is changing the polling cycles to 60 seconds and see if you still have the same "peaking" issues.

It might be that you're trying to poll too often, you're getting answers late at those 30 minutes periods and it skews the results a bit.

G.

[poskalp]

Hi gleepwurp, thanks for your answer.

I'm querying the device every 20 seconds because it is a border router and we need to know what is happen with the maximun posible detail.

I going to try your advice, if this is the cause, there is any option or pluging for zabbix to poll very often but process the response every minute, for example, something like the boost option in Cacti?

Thanks in advance for your answer