We have a customer with three different brands of networking devices spread across ten subnets. Each brand has it's own template. Devices are added/replaced often enough that it makes sense to automate discovery. Since each brand of device has it's own template and the devices are spread across subnets here is what I've done. It is working but I'm hoping there is a better way.
Separated each subnet into it's own Discovery rule (10 rules). This is so they can run in parallel.
Each rule has four SNMP checks; one for the sysName which is used as the Host Name, the other three allow me to tell what kind of device it is.
Created three Discovery Actions; one for each brand device. Each action has ten conditions; one for each of the discovery rule (subnet) checks that corresponds to its device type.
This is not very scalable. If I add another subnet I need another Discovery rule and then have to edit all three actions and add yet another OR condition to match it. It would make more sense to have the ability to do filtering/checks on the action but that doesn't seem possible. It would also be nice to ping devices first and then do SNMP checks if there is a response; zabbix seems to waste a lot of time waiting on SNMP timeouts.
Any recommendations?
Separated each subnet into it's own Discovery rule (10 rules). This is so they can run in parallel.
Each rule has four SNMP checks; one for the sysName which is used as the Host Name, the other three allow me to tell what kind of device it is.
Created three Discovery Actions; one for each brand device. Each action has ten conditions; one for each of the discovery rule (subnet) checks that corresponds to its device type.
This is not very scalable. If I add another subnet I need another Discovery rule and then have to edit all three actions and add yet another OR condition to match it. It would make more sense to have the ability to do filtering/checks on the action but that doesn't seem possible. It would also be nice to ping devices first and then do SNMP checks if there is a response; zabbix seems to waste a lot of time waiting on SNMP timeouts.
Any recommendations?
Comment