There are 3 ways to selectively enable port monitoring through polling. These will regularly check the mib for ifOperStatus and ifAdminStatus of the port. If the ifAdminStatus is up, and the ifOperStatus is down, the port will be considered to down.
PollPortStatus. This is enabled by default at device level, but disabled per port level so as not to induce too much load on the SPectroSERVER. To have a port get polled, you need to have both the device and port's PollPortStatus=True. If this is not still working as desired, then check:
polling interval is not = 0
isManaged = true
Live Pipes. When these are enabled by right clicking on a resolved pipe, SPECTRUM will set ok_to_poll=true on each end port. This will enable polling on those ports. Verify that this is not disabled globally in the Landscape>Configuration view.
WA_link. Can set to ports only, link only, or bothportsandlink. Check to ensure the WA_Segment models are properly resolved over the appropriate end ports
In the VNM Component Detail View> Configuration > Fault Isolation you can set Port Fault Correlation to Disabled, All Connected Ports, or Management Neighbors Only. This covered in depth in the How to Manage Your Network with SPECTRUM document under Configuring for Fault Isolation.
What are the ways to enable poll-based monitoring of port status?
I have configured to have port level polling; what can I check if a port is not being managed as expected?
Also reference TS30278 Monitoring Port Status using Traps in SPECTRUM
(Legacy KB ID CNC TS30276 )