You are here

Need clarification on notification logic

4 posts / 0 new
Last post
Karl_MCS's picture
Need clarification on notification logic

I've been using Opsview for a while, but never really got a firm grasp of the logic behind the notification profiles.  So you have hosts, services, and keywords.  I'm aware of the bug where selecting all hosts but no services or keywords will result in no host alerts.  And I get that if you select some hosts and all services, you'll get all host and service alerts only on the selected hosts.

What I've been trying to get a handle on is the keywords.  They can be on both hosts and services.  The notification filtering logic is completely undocumented in the wiki.

Are they inclusive, meaning that notifications will be sent for a given selected keyword, even if the host or service group itself is not selected?

Are they exclusive, meaning that notifications for selected host and service groups will be sent only if their keyword is also selected?

Or something else entirely?  Please clarify.  And also, does this same logic apply on the role permissions?


smarsh's picture
Re: Need clarification on notification logic

Basically, you need to think of it as 2 seperate sections (we changed this recently but im not sure it got pushed into Core). 

You have: hosts/host groups and services - here you select, "i want to be alerted for critical/warning/recovery/... on these hosts".

Or alternatively, you have keywords. Here you create a keyword 'Notifications-123', and then go into each service check or host and add this check. Then, when a service check errors that has this keyword attached, then the alert will be sent.



PS: If you are struggling to understand Opsview, i'd recommend one of the Opsview training courses as they can run through this.

Karl_MCS's picture
Re: Need clarification on notification logic

Ok, so essentially what you're saying is the host/service group checkboxes and the keyword checkboxes operate independently of each other.  Is that pretty much it?

So I'm still left with some questions about the keyword logic.  Let me try to come up with an example.  Let's say we have:

  1. Host-1: keyword is on host but not on any of its services.
  2. Host-2: keyword is on host and one of its services.
  3. Host-3: keyword is on a service but not on the host itself.
  4. I am configured to receive email alerts for this keyword and nothing else.

Now let's say...

  • All 3 hosts go into a hard down state.
    • Do I receive host down alerts for Host-1 and Host-2?
    • Or only Host-2 because Host-1 doesn't have a service with that keyword?
  • All 3 hosts are up, but the service with the keyword goes hard down.
    • Do I receive service alerts for Host-2 and Host-3?
    • Or only Host-2 since the keyword is not on Host-3?
  • Any selected/unselected host/service groups have absolutely no bearing on the above other than adding additional things to be notified about?

Thanks for your help.

smarsh's picture
Re: Need clarification on notification logic

Correct- they are indepedent of each other.