cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 9

Best practice for handling nodes intended to be down normally in NPM?

Jump to solution

Here's my situation:

One of our clients have implemented backup internet circuits that are normally programmed to be down, as they are metered internet circuits that charge by the gigabyte. However, if the main internet circuit goes down, the router is programmed to bring the metered circuit up. The client wants alerts for when these circuits come online.

I have setup an alert for when one of these nodes go up, and made them differentiateable from other nodes via custom property. That part was easy. (For reference: Node's normal condition is down, how to deal with that in NPM ) My question is there's anyway to make these nodes visually distinct from other nodes? They are suppose to be down, and it can be confusing to my teams to see a red down node icon for a down node that is normally supposed to be down instead of up.

Ideally I would like these nodes to appear different from the others, like a different icon set that looks different from most other nodes. Is there a known way to accomplish this? Searching on Thwack I come across a lot of articles for using custom icons in Network Atlas, but I had trouble finding anything for NPM.

Labels (2)
Tags (5)
0 Kudos
1 Solution
Level 16

Hi

Juniper has red/yellow alert

try to see if you can get those interface uplink  status to red yellow

That will indicate to your noc at there is a problem

Understanding Alarms - Technical Documentation - Support - Juniper Networks

View solution in original post

7 Replies
Level 16

Hi

Juniper has red/yellow alert

try to see if you can get those interface uplink  status to red yellow

That will indicate to your noc at there is a problem

Understanding Alarms - Technical Documentation - Support - Juniper Networks

View solution in original post

I found an OID that can be used to monitor JunOS Probes: 1.3.6.1.2.1.80.1.3.1.7 (listed as pingResultsProbeResponses)

With that said, to answer my own question, every node we've setup that's normally down but is up in case of a failover event has been named to [FAILOVER] (Rest of Node Name). The idea being that if one of my team members sees [FAILOVER] at the start of a node name they'll know it should be normally down, and to take action if it's up.

0 Kudos

Cool

is that just a universal poller or you could  add the RPM service failure to red yellow?

sja

0 Kudos

It is a universal poller that I added to watch the probes, as that's what triggers the ip-monitoring policy on the Juniper router to turn on the interface to the metered internet circuit.

To make it useful I also added a 2nd universal poller for readable labels 1.3.6.1.4.1.2636.3.7.1.3.1.22 (jnxPingResultsOwnerIndex). Once both are assigned to a node I get a nice table that looks like this:

2017-02-21 15_09_16-Node Details - Summary - MI-10121-RTR01.png

MVP
MVP

If you set them to be external there won't be any polling (therefor no green/ red/ yellow status icons).  You can always change that back when you need to.

0 Kudos

When I had to do the same thing, them in a separate group. Then leveraged custom alerting to make them alert backwards. Are you using Cisco IP SLA type functionality?

The setup is similar to Cisco IP SLA. We're actually using Juniper ip monitoring. Juniper Networks - [SRX] Example - IP Monitoring with route fail-over configuration and behavior - K...

0 Kudos