This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Best way to monitor switch interfaces

How is everyone doing this? Ideally, you'd want to monitor all of your interfaces, but not alarm on them if they're not in use. You can set the interface to admin down on the switch, and Orion then knows it's not down, but then if you use that interface in the future you will have to make sure it's checked in Orion for monitoring. This seems like an extra step, and can result in critical interfaces not being monitored if someone forgets or misses that step. Also to consider are ports that change dynamically (IE; a port a user is connected to goes down because they've shut off their computer). Is there a way to do this without the manual work of constantly unchecking/checking interfaces you care about?

  • Assuming you have a large enough license you would just want to monitor all interfaces and everything that isn't an uplink or permanent connection should have the flag for "unpluggable" set to true.  Now you have full visibility and only get notified on down interfaces for the ones you hadn't tagged as unpluggable.  The hardest part then is the leg work of figuring out which interfaces should be permanently up.

  • We tag all interfaces with 'unpluggable' unless we want to alarm on that interface or it's being specially monitored.  This really keeps the noise level down and keeps Solarwinds clean.

    However, we go one step further to prevent accidental down interface notifications.  We incorporated a interface custom property and we then built alarms around it.  The beauty in doing this is we quickly and easily find the interfaces that send out down/notifications and we can build reports around it. 

    I'll admit it was a lot of work to setup and to train everyone, but once it was in place it was a god-send.

  • I agree with the other responses. In your case setting it to unplugged seems to be the best option. This way you don't have to remove re-add constantly. No need to have to keep rechecking things, etc. You monitor them. Chose the ones that aren't critical set unplugged to try and only get alerted when critical interfaces alert.

  • This is an interesting method, but even in this scenario if we take a previously "unpluggable"  interface and then use it for something critical someone still has to go and change the status of that port in Orion right? Also, where is this set at?

  • That is correct.  You would make these changes under interface properties.

    In regards to you "what if" scenario, that comes down to standards, definitions, rules and regulations.  At the end of the day their should be polices written and enforced on "critical" connections along with what should and shouldn't be alarmed on. Case in point someone shouldn't be able walk into your environment and have the ability to unplug a critical connection like a DHCP server. Assuming that is the case the next logical question would be, where is the alarm generating from, the down interface or the node down?  That is of course assuming everything is being managed in Solarwinds.

  • A little while back, I submitted a feature request to automatically manage switch ports connected to a managed device.  I don't monitor printers or PCs...so managing all switch ports put unnecessary load on my pollers.

    Please vote it UP!

  • I never monitor all interfaces on switches. Please consider to define (really) critical interfaces - those which have to be always UP. Other words if critical interface goes DOWN it means something wrong happened. Even if you have SLX license it is very little information that interface with workstation hooked goes down. I mean who cares about that? Maybe user unplugged the cable or mabe laptop has been disconnected for other reasons. Anyway if you monitor non critical interfaces than switch shows up tiny red square blinking what means that there is something wrong with it - bit messy isn't it? In general please consider what is valuable information from interfaces being monitored on your switches. Similar situation for servers - if you monitor ports on switch with server LAN interfaces than it is not necessary to monitor interfaces on server.

    I agree with bourlis - all scenarios depends on your standards, definitions etc. but first think about sense of monitoring that particulat interface on switch - is anybody who needs the information?

  • t0ta11ed74​,

    Based on your responses above, it appears you are looking for an automated way to detect "important" ports from regular user ports; this way you can automatically monitor them differently. To my knowledge, there is no such automated mechanism (out of the box). Further, "most" folks only monitor important ports with NPM and don't really bother with regular ports (based on my experiences); User Device Tracker (UDT) was specifically created for user/regular ports.

    Regardless of the monitoring methodology you choose (important only, or important with others as unplugged), it is still a manual process to change the monitoring method of an interface, it becomes an administrative task in your process.

    D

  • Is there a way to find all the "down" ports and make them unpluggable?