You really want a performance template to do this sort of frequent monitoring. If you REALLY need to check things, say, every minute (and I wouldn't go more frequent than that) then you probably need to use a command template as you can then control the cycle time, whereas with an SNMP template you may be back to a single zenperfsnmp daemon with a single polling interval, typically of 5 minutes. If you have 3.2.1 then you can apply an SNMP polling interval per device which gives you lots more flexibility.
It also depends on how many of these devices you have. If you use a command template then it is a performance hog so don't go this way unless you only have a few devices you need to monitor.
The next question is what you mean by "ports up or down". What is this device? Does it have an SNMP OID that you can monitor that gives port status? If it is a switch then most devices do provide an enterprise-specific OID for port status or you may get it out of the Bridge MIB if the device supports that. If you are simply talking interface monitoring on servers, then you can use the standard ifOperStatus from MIB-2 - this is what is used by zenmodeler to determine status. It is perfectly possible to use the same variable in a performance template and then have a threshold on >1 (ifOperStatus is defined in RFC1213 MIB as 1=up, 2=down, 3=testing). Have a look at this item - docs/DOC-2494 .
Cheers,
Jane
