SNMP Timeout > 60 sec – multiple devices affected

Troubleshooting
,
1

Hi everyone,

I’m running Checkmk CEE 2.4.0p29 and experiencing sporadic SNMP timeouts on multiple devices. The SNMP fetcher occasionally exceeds the 60-second limit:

Copy[snmp] Fetcher for host "..." timed out after 60 seconds
CRIT – Missing monitoring data for all plugins

Environment

  • Checkmk CEE 2.4.0p29

  • Affected devices: Various SNMP devices (management interfaces, switches, etc.)

  • Network: Some devices connected via VPN / branch office links but the most are in our local Network.

What I have already tried

  • Adjusted Timing settings for SNMP access (increased timeout & retries)

  • Checked firewall rules

  • Tested SNMP connectivity via snmpwalk from the Checkmk server – works fine

Question

Is there a way to permanently raise the 60-second hard timeout for specific hosts? Has anyone dealt with slow SNMP devices and found optimal settings for this?
I’m not safe that the problem really are slow SNMP Devices. I actually don’t think thats the problem.

Thanks!

2

Hi @mk_fut,

Helpfully is it to use the rule Hosts using a specific SNMP Backend (Checkmk Enterprise only) and set it to classic backend. Only for this host, with the problems.

After this you can try to debug it more, by figure out, what OID would be so slow.

To do this, you just can use this command, as site user on the CLI of you checkmk server, to see, at which OID checkmk will stop:

OMD[mysite]:~$ cmk -IIv HOSTNAME

After the walk stops or hang at a specific OID, just use this command to figure out at which bulk size it will be working:

OMD[mysite]:~$ snmpbulkwalk -Cr<BULK_SIZE> -v<SNMP_VERSION> -t <TIMEOUT> -c <COMMUNITY> -m "" -M "" -Cc -OQ -OU -On -Ot <IP_ADDRESS> <OID>

Here is a short overview about the parameters:

BULK_SIZE – Values to test: 8, 10, 24, 60
-v - SNMP version (2c or 3) example: 2c
TIMEOUT – Timeout in seconds per OID: 1-7

If you found the correct value for the bulk size, you can use this value in this rule Number of OIDs per bulk. But please only for this one host:

image
image986×1037 72.3 KB