Titus Anderson wrote:
> I really don't have a use for this idea, yet, but I thought I'd throw it out
> here and see what people think. (I also don't know how hard this would be to
> code -- I haven't dug that deeply into the scheduling code, yet.)
>
> Here's a theoretical setup. Say I have a number of hosts that all provide LDAP
> services and also have other services monitored (it doesn't matter what, but
> there need to be other services for this to be relevant). With interleaved
> checks, the point is to spread out checks across multiple hosts. If the
> interleaving factor works out exactly the right way, it's very possible that
> every one of the LDAP services could be checked simultaneously. For redundant
> services, that's not necessarily desired behavior.
>
> So, here's my idea. Since we have servicegroups now, perhaps providing a
> scheduling hint would be possible. If I have 5 redundant LDAP services and I
> want at least one of those checked every 5 minutes, to guarantee that now, I'd
> have to set the check_interval to 5 minutes. However, if I could hint to the
> scheduler to always interleave the checks of those services, I could
> comfortably set the check_interval to between 20 and 25 minutes.
>
> Of course, there's also the flip side -- you may have a set of services that
> are related but not redundant that you want to always check together. So it
> would be better to parallelize those checks.
>
> So, finally the point -- the idea is to add an option to servicegroups (and
> possibly hostgroups for those that use scheduled host checks) that provides a
> hint to the scheduler. Something like scheduling_hint with options of
> "interleave" or "parallelize".
>
> Any thoughts? Anyone have an idea of how difficult this would be to implement?
> Anyone find this useful or a complete waste of time and effort?
>
In my opinion, implementing stuff that you don't have a use for is a
waste of time. Better then to implement in-core loadbalancing/redundancy
features, which a lot of people would most likely use. I have the
schematics for it ready, so let me know if you want to plug in on it.
--
Andreas Ericsson [email protected]
OP5 AB www.op5.se
Lead Developer
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]