SNMP checks in discovery rules

[cerasus]

It appears that if a discovery rule contains several SNMP checks, each one will be executed in sequence. If there's really no device on the other end, the first check will time out after 3 seconds (by default), followed by another check that will take the same amount of time and so on. So, if there's no device on the other end, it will take 3 * number of checks seconds per IP.

Is there a way to abort all SNMP checks if nothing is listening on port 161? Maybe based on an ICMP ping against the IP?

[Hamardaban]

https://www.zabbix.com/documentation/current/manual/appendix/items/unreachability

[cerasus]

Thank you for your reply. I don't think we're talking about the same thing though. I'm talking about reachability during network discovery. The SNMP checks against dead IPs take too long and as a result the whole discovery process is protracted unnecessarily long.

So I'm wondering if there's a way to short circuit the defined SNMP checks in the discovery rule IF the IP doesn't respond to a ping.

[cerasus]

Actually what I'm noticing is that with my Timeout configured to be 3 seconds and two discovery checks, it takes 6 seconds per check or 12 seconds per IP

[cerasus]

I think I may have found the answer: https://www.zabbix.com/documentation...itemtypes/snmp

Since Zabbix 2.2.8 and Zabbix 2.4.2 Zabbix server and proxy daemons will always retry at least one time: either through the SNMP library's retrying mechanism or through the internal bulk processing mechanism.