I have read many posts on the topic, but there still does not seem to be an useable answer.

Problem:  Opmanager generates alarms for any switch ports that are down on a managed device. 

Scenario:  The main problem is having alarms on access switch ports. Alerts on routers or core-switches are a good idea, but on access switches, alarms are a nuisance. I have approx 2400 access ports, and with say 10% of those spare, there are a lot of alarms.

Disabling polling or management means we lose out on valuable information.

We still want to have: 

• Port status (so we can easily see port utilisation)
• Get alerts on errors/discards
• THroughput
• etc.

The key issue for me, is that on an Access switch, an access port down status is not a fault.

But a trunk port down, is a fault.

Eg.  I use 2960's as access switches, and 3750's as core switches.

On the 2960 I use, there are 24 FA ports, and 2 GI ports.

The FA ports are access ports, and i dont want an alarm generated for their link status. 

The GI ports are used to link to other access switches, or back to the core switch. I DO  want alarms for link down alerts on these ports.

On the 3750's, all ports are used to link to servers, routers or other switches, and I want alarms for link down.

Solution:

What we need to be able to do, is control whether link status generates alarms, for a given interface.

Ideally, we could set these preferences on the device template, so that it applied to every switch.

I would happily swap  the "Poll Status" option for " Alarm for Status".