Welcome to Questions - Ask OpenNMS, where you can ask questions and receive answers from other members of the community.

How do I set OpenNMS to have two types of Node Down events with different severities?

0 votes
45 views
For example, I have laptops in one subnet which keep turning on and off so if a node on that subnet went down I'd class that as an intermediate incident. However, on another subnet there are critical servers which if one went down I'd class as a major incident. How can I get OpenNMS to treat it separately, preferably by different events but different alarms would be a viable option?
OpenNMS version
20.0.1
Java version
java version "1.8.0_131"
Java(TM) SE Runtime Environment (build 1.8.0_131-b11)
Java HotSpot(TM) 64-Bit Server VM (build 25.131-b11, mixed mode)
Operating system
Windows Server 2016
PostgreSQL version
9.6.3
asked Jul 31 by louis63 (200 points)

1 Answer

+1 vote
 
Best answer

Without more details about how you interact with OpenNMS, it's hard to say what would be the best approach to achieve your ultimate goal. I can say, however, that trying to have "different types" of nodeDown events sounds like the wrong approach. You could probably do it, but you would be in for a long struggle against the design of the poller.

If you don't care at all about the up/down status of the laptops, you can just tell the poller to exclude them from its normal operations. A great way to do this is with a surveillance category used in the filter expression of a poller package. For example, you might put all the nodes whose up/down state you do care about into a category called Track_Uptime. Assuming you're using the default version of poller-configuration.xml you might then change the example1 package like so:

  <package name="example1">
    <filter>catincTrack_Uptime</filter>

    ...

(That's not a typo, there's really no space between "catinc" and the category name. The syntax is a bit unwieldy.)

Because of a known bug in 20.0.1, you'll probably need to restart for this change to take effect. Now only services on nodes belonging to that category will be monitored for up/down state by the services defined in that package. Other nodes can go up and down all they want and the poller will say nothing about it.

I realize this may go too far in the opposite direction for your liking. OpenNMS is flexible enough that a million different in-between solutions are achievable; the above is just one possibility that lends itself to concise summary. I hope it's helpful.

answered Aug 3 by jeffg (760 points)
selected Aug 4 by louis63
Hi Jeffg,

Thank you for answering my question. I think your solution will work for my needs and I appreciate you pointing me to a better way of solving this issue rather than trying force this idea to work.

Many thanks,

Louis
...