Passive service checks via NSCA and requir ed check command defini tion

[Ralph.Grothe at itdz-berlin.de]

> > I missed this thought before sending my last reply. 
> > 
> > Would you suggest I enabled active checks for the service
> 
> No.  If you do that then both active and passive checks will
occur on
> that service.  You need to disable active checks on any service
that
> accepts passive check results.  The active check is triggered
if the
> passive check is stale.  Read the documentation because there 
> are other
> things you need to get right for staleness checking to happen. 

You got me, I so far skipped the chapter on staleness checking
because I considered it to be an advanced topic, and I'm still 
in my Nagios initiation/infancy.
But one runs pretty soon into those topics if one wants to do
some
"real world" monitoring.
I will revisit the docs tonight.


> 
> > and provided the dummy_check returning state UNKNOWN
> > in order to get these states whenever the passive checks
didn't
> > take place, for whatever reason?
> 
> Yes. 
> 
> > Well the only reason I can think off is that the remote host
> > crashed (or the route thereto).
> 
> I can think of other possibilities.  Crond is started late in
the init
> sequence, after network services.  If there is a problem with 
> something
> prior to crond starting crond may not get started but 
> everything you're
> monitoring on that server is started. 

You're absolutely right.
It was a bit short sighted by me.


> 
> > But I covered this already by host icmp checks.
> 
> Which won't occur if all the network services are running but 
> crond isn't.
> I've also seen machines that run out of resources able to
respond to
> pings but nothing else. 

Beleive it or not, I myself already experienced that a server
was,
what they call in memory management terms, "thrashing".
The application (database monster with probably ill-fit queries,
indexes, params etc.)
was consuming so much memory (viz. desperation swapping) that
even crond
demised.
I surely will have to set up a monitor for things like that.

> 
> If whatever it is you're monitoring is on that same machine 
> then you'll
> probably know there's a problem because its network services go
down.
> But if it's a firewall submitting passive check results about
machines
> behind the firewall (for scaleability or because those 
> machines are not
> reachable from the outside) then the icmp checks are not
necessarily
> going to tell you that there really is a problem. 

Ah firewalls, another of my monitoring intentions oponents.
Even worse with ours.
Firewall admins seem (for paranoia reasons?) to silently drop all

ICMP packets from even ingress (no not the database ;-) traffic,
what to me seems counterproductive to ICMP's original objective.


> 
> > Another thing strikes me if I enabled active checks. 
> > 
> > How should I space them, meaning how to size the intervals?
> 
> The interval doesn't matter because you don't enable active
checks. 
> 
> > Maybe my thinking is a bit far-fetched?
> 
> I think you need to read the docs some more.  Unfortunately the
> information you want is scattered around several different
pages.


very scattered and disjoint indeed
 


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null