Please Alexei... help me...

[llopht]

Hi,

I use this item for all my hosts for the outgoing traffic :

Type : SNMPv1 agent
SNMP OID : .1.3.6.1.2.1.2.2.1.10.16777220
SNMP Port : 161
Key : TrafficOUT
Units : bits
Use multiplier : do not use
Update interval : 30
keep history : 7
keey trends : 365
Status : monitored
Type of information : numeric
Store value : Delta (simple change)

I have two problems :

1. Values are incorrect (ex. One of my host can use a bandwidth with a maixmum of 1Mbytes, I got a 4Mbytes value with zabbix). Mrtg give me the correct value with this OID.

2. When I restart HTTPD I have a big peak in my picture (see http://www.actisens.com/screen.gif)

Please help me...

Jérôme

[Wolfgang]

Hello,

You should have a look at the raw value collected via zabbix. Do they meet your expectation?

latest data -> select host -> history graph -> values of last hour

Also, can you try changing:
-Store value : to "As is"

And/Or change to different values:
-Units :
and eventually (but this should not be necessary)
-Use multiplier :


Btw: Next time i would choose a different topic name like:
"Zabbix Agent returns wrong values via SNMP" and also add what version you are using. That might enhance the number of responses you will get.

[cadbury]

to have the correct values, you might store values as
delta (speed per second)
because with delta (simple change) zabbix just do the difference, and you have the number of bytes (not bits), that INcome (the .1.3.6.1.2.1.2.2.1.10 mib is .iso.org.dod.internet.mgmt.mib-2.interfaces.ifTable.ifEntry.ifInOctets) during the time between the current and the last check (30 seconds here).
so with delta (speed per second) zabbix will make the difference between the last and the current value, and divide it by the time elapsed, so it make the average speed during this time.

for the peak, it's happening only when you restart the httpd deamon?
are the values around this peak corrects?

[cadbury]

after thinking, the peak is also due to the way you store the values ( delta (simple change) ).
on your graphs, no values are retrieved from a long time, but the mib value is still increasing (this value is the total number of bytes recieved since the network interface was last started). so when the first value is retrieved, there is a great difference, and that make the peak.
with delta (speed per second), you won't have this problem

[Alexei]

after thinking, the peak is also due to the way you store the values ( delta (simple change) ).
on your graphs, no values are retrieved from a long time, but the mib value is still increasing (this value is the total number of bytes recieved since the network interface was last started). so when the first value is retrieved, there is a great difference, and that make the peak.
with delta (speed per second), you won't have this problem

Absolutely correct!