RE: [Nagios-devel] service definitions
Posted: Wed Mar 26, 2003 1:49 pm
(Preamble: Forgive me for anything I may utter in this e-mail, as last
night I was working into the wee hours of the morning, and am running on
about 3-4 hours' sleep.)
Matt McKinnon wrote:
>
> in the nagios documentation there are "time-saving tricks"
> you can use. one
> example is:
>
> define service{
> host_name *
> service_description SOMESERVICE
> other service directives ...
> }
*nods* I'm familiar with this feature.
> let's use that as our template for a very simple service
> definition that
> uses the command "check-host-alive" to ping all hosts [*]
> across the board.
> that's excellent, because i can add one thousand hosts and
> that service
> check will automatically include them all. that IS time saving!
Service definition? Or host definition?
I don't mean to quibble, but check-host-alive is generally used in the host
definition. (This might not even matter in the context of what you're
trying to describe to me. YMMV.)
> now let's pretend that i have a couple contracts that i'm
> monitoring, each
> of them with a different contact address. if a host goes
> down on one of
> those contracts, users should have the option of letting
> nagios figure out
> who to send mail to (based on obvious rules), instead of
> users needing to
> have one "contact_groups" directive set.
My apologies for being picky, but... is it safe to assume this:
s/contracts/services/g
Otherwise, I believe I'm following....
> >Assuming there were no requirement for a contact group, then
> there would be
> >no notifications.
>
> that shouldn't be the case.
When you say "shouldn't", do you mean, "this isn't the logic that Nagios
currently has implemented, therefore it shouldn't behave this way", or "I
understand that this is how Nagios behaves today, but I would rather it
behaved this other way"?
> as i said before, nagios should
> know that:
See my last comment. (s/shouldn't/should/g)
> . that host [which has gone down] is part of a host group
> . that host group has a contact group
> . that contact group has a contact
> . that contact should receive mail if host goes down and no
> contract_groups
> are set in the service definition.
I don't believe that functionality exists today. This appears to be a
reasonable RFE. I might even consider leveraging that functionality myself
here and there, to save a few keystrokes.
> did i lose you in the dark, or did i shine some light down?
I *think* I'm seeing the light; all that remains is for me to cut through
the fog of sleep deprivation. :/
> -matt
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
night I was working into the wee hours of the morning, and am running on
about 3-4 hours' sleep.)
Matt McKinnon wrote:
>
> in the nagios documentation there are "time-saving tricks"
> you can use. one
> example is:
>
> define service{
> host_name *
> service_description SOMESERVICE
> other service directives ...
> }
*nods* I'm familiar with this feature.
> let's use that as our template for a very simple service
> definition that
> uses the command "check-host-alive" to ping all hosts [*]
> across the board.
> that's excellent, because i can add one thousand hosts and
> that service
> check will automatically include them all. that IS time saving!
Service definition? Or host definition?
I don't mean to quibble, but check-host-alive is generally used in the host
definition. (This might not even matter in the context of what you're
trying to describe to me. YMMV.)
> now let's pretend that i have a couple contracts that i'm
> monitoring, each
> of them with a different contact address. if a host goes
> down on one of
> those contracts, users should have the option of letting
> nagios figure out
> who to send mail to (based on obvious rules), instead of
> users needing to
> have one "contact_groups" directive set.
My apologies for being picky, but... is it safe to assume this:
s/contracts/services/g
Otherwise, I believe I'm following....
> >Assuming there were no requirement for a contact group, then
> there would be
> >no notifications.
>
> that shouldn't be the case.
When you say "shouldn't", do you mean, "this isn't the logic that Nagios
currently has implemented, therefore it shouldn't behave this way", or "I
understand that this is how Nagios behaves today, but I would rather it
behaved this other way"?
> as i said before, nagios should
> know that:
See my last comment. (s/shouldn't/should/g)
> . that host [which has gone down] is part of a host group
> . that host group has a contact group
> . that contact group has a contact
> . that contact should receive mail if host goes down and no
> contract_groups
> are set in the service definition.
I don't believe that functionality exists today. This appears to be a
reasonable RFE. I might even consider leveraging that functionality myself
here and there, to save a few keystrokes.
> did i lose you in the dark, or did i shine some light down?
I *think* I'm seeing the light; all that remains is for me to cut through
the fog of sleep deprivation. :/
> -matt
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]