Item not found in monitoring data

We are having some problems like UNKN status for some Interface checks which are working via SNMP from ESXi hosts. We have the summary message “Item not found in monitoring data”.

Yesterday we had 8 UNKN status for some network interfaces (some vethernet#### and vfc###) simultaneously – four interfaces at each host, at same time per host.

Today we got another 8 UNKN status checks for another two hosts appeared today, also at same time per hosts.

All services (including those vethernet and vfc) are active. The manual SNMP requests returned ifOperStatus for each of the interfaces are returned. SNMP ifTable also returned, nothing suspicious there.

We are running CheckMK ver. 2.0.0p6.

Almost exact situation as here - CMK 2.0.0p7 if64 - UNKOWN with "Item not found in monitoring data"

Is it a bug or something that is reported to be fixed?

I know that SNMP can be “hit or miss” (tends to timeout or not respond on hosts). While we’re just doing same for piggyback data from the DRAC (Dell hosts), the DRACs stop repsonding everyday, unfortunately, the nodes aren’t in sync. Maybe at some point, I’ll write specific “time” rules for each, but anytime a node goes into maint and reboot, the time will be different when the DRACs go on their “coffee break”.

With that said, we have other devices that have difficulty with SNMP, like printers and heavily stacked switches. In the switch case, I just disabled the notifications. In the printer case, I’m trying Status of the Check_MK services, rules… but we’ll see.

Beside the problem @cjcox already mentioned i would say

is a bad combination. If you query the ESX host with the special agent you should get the traffic of the hardware interfaces and this traffic should also be valid.

This is correct “vethernet##” and “vfc##” are virtual interfaces only existing temporally. If you move one virtual machine to another host some of the interfaces will vanish and if you create some virtal machine you get new ones.

I would prefer to switch of the SNMP on the ESX host and only monitor it with the vSphere API.

1 Like

Thank you very much for your answers, @cjcox & @andreas-doehler

I do see the “timeout” sometimes when I make snmpwalk query to the ESXi server manually.

Also, I would exclude vethernet and vfc interfaces from being checked. Also we’ll try to use vSphere API as well.

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.