Re: [Nagios-devel] RFC/RFP: Service parents
Posted: Wed May 18, 2011 9:09 am
On 05/17/2011 02:14 PM, lists wrote:
> You can probally ask these questions your self but...
> If my understanding of this is correct, How will this work in a
> templated environment?
> eg Will the parent service be configured via a hostname / service name pair
> Or just a service name, that Nagios will look on the applied host to find?
> (I ask as we use host templates to apply service to hosts).
>
> I am probably being very anal here, would this situation work?
> Host
> |
> - service0 NRPE -->
> -- service1 App check1
> -- service2 App check2
> |
> - service3 HTTP check1
> - service4 HTTP check2
>
> Could we have "HTTP check1" depend on "App check1" but not "NRPE"?
>
Yes, but if NRPE is a real parent (ie, a binary check to see that
the agent is running), there's no way it will fail without also
making App check1 failing.
A bigger question though is why you'd want to. If something serves
the proper html page over the network, does it really matter to you
if App checkX is working?
> eg NRPE dies so service1 alerts are suppressed, but "service3" will
> alert if needed.
> But if "App check1" fails (but not NRPE) "HTTP check1" will not alert?
I should think so, but that depends on how we decide to implement it.
A service that returns OK when its parent has returned non-OK is
obviously misconfigured (since it really shouldn't be ok with a non-OK
parent), but there's no real reason to dive deeper than a single level
for parenting checks. We can just let users shoot themselves in the
foot with misconfigured parent chains and ignore all multi-depth
traversals completely.
--
Andreas Ericsson [email protected]
OP5 AB www.op5.se
Tel: +46 8-230225 Fax: +46 8-230231
Considering the successes of the wars on alcohol, poverty, drugs and
terror, I think we should give some serious thought to declaring war
on peace.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
> You can probally ask these questions your self but...
> If my understanding of this is correct, How will this work in a
> templated environment?
> eg Will the parent service be configured via a hostname / service name pair
> Or just a service name, that Nagios will look on the applied host to find?
> (I ask as we use host templates to apply service to hosts).
>
> I am probably being very anal here, would this situation work?
> Host
> |
> - service0 NRPE -->
> -- service1 App check1
> -- service2 App check2
> |
> - service3 HTTP check1
> - service4 HTTP check2
>
> Could we have "HTTP check1" depend on "App check1" but not "NRPE"?
>
Yes, but if NRPE is a real parent (ie, a binary check to see that
the agent is running), there's no way it will fail without also
making App check1 failing.
A bigger question though is why you'd want to. If something serves
the proper html page over the network, does it really matter to you
if App checkX is working?
> eg NRPE dies so service1 alerts are suppressed, but "service3" will
> alert if needed.
> But if "App check1" fails (but not NRPE) "HTTP check1" will not alert?
I should think so, but that depends on how we decide to implement it.
A service that returns OK when its parent has returned non-OK is
obviously misconfigured (since it really shouldn't be ok with a non-OK
parent), but there's no real reason to dive deeper than a single level
for parenting checks. We can just let users shoot themselves in the
foot with misconfigured parent chains and ignore all multi-depth
traversals completely.
--
Andreas Ericsson [email protected]
OP5 AB www.op5.se
Tel: +46 8-230225 Fax: +46 8-230231
Considering the successes of the wars on alcohol, poverty, drugs and
terror, I think we should give some serious thought to declaring war
on peace.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]