Closed

Closed due to inactivity. Received 10 votes with last vote on 11 Nov 2019.

Options for unwanted Interface Down alerts post-discovery

We have a bit of a problem with NPM in our environment as it relates to adding (or discovering devices).  In our data centers, all interfaces are added to monitoring -- regardless of their status.  This, so that SW can monitor and report on any changes and/or problems with each of the interfaces.  Now, to save ourselves from getting too many alerts, all ports on data center switches (except the Uplinks) are set as Unplugged.   However, the problem comes when the new node added has some interfaces many interfaces (over 300) and some of those interfaces are currently down.  By the time I get to edit the interfaces and flag those as Unplugged, some alerts always make their way to our ticketing system.   This is far from optimal and causes confusion and frustration to our end users and us (as SW admins) too.

So, now that I've given you a picture of the problem, here are the two options I came up with --- either of this should do the tick but I favor #1:

  1. Allow NPM to flag interfaces as Unplugged as part of the discovery process.  That way, NPM "already knows" which newly added interfaces should be alerted upon.
  2. Have NPM allow a delay for newly added devices (1 hour???), so the SW admins have plenty of time to flag interfaces as Unplugged (if they want to).


What about you? Do you think that this delay in alerting a good idea?  If so, I would greatly appreciate your vote.  emoticons_cool.png

PS:  I would also very much like to hear from other SW admins facing similar issues as me.  What workarounds (if any) have you implemented to avert getting alerts on interfaces you don't want to alert upon?