Reports only show data from a specific tim e period?

[Paul L. Allen]

Matt Pounsett writes: 

> On Thu, 5 Feb 2004, Paul L. Allen wrote:

>> The CGIs display more than statistical information.  And even assuming
>> they are only interested in the statistical information, suppressing
>> raw data turns statistics into a LIE.  It is known as selective sampling.
> 
> It's not a lie if that's the statistical information you're asked for.  
> "We have had no outages between 7am and 7pm weekdays," is made a lie 
> because the webserver was out for five minutes on Saturday.

So let's examine that closely.  Either there are significant outages
outside of hours or there are not.  If there are no significant outages
outside of hours then the full data can be processed and it will still
look healthy.  If there are significant outages outside of hours that is
a sign of system instability or impending failure but the statistics
will show "no problem" and that is not a lie, according to you. 

Let me think about this.  Why would somebody want to know if machines
are failing outside of working hours but not let the customer know about
that?  One reason might be so that the machines could be fixed on the
pretext of "preventative maintenance" without the customer realizing they'd
been sold unreliable rubbish.  Remember that if there are no significant
problems outside of working hours then the statistics will show good
health anyway and there would be no need to suppress data. 

> He's not requesting that the entire system only report data between those
> hours -- he's asking for trends and graphs that show statistics for those
> hours.

If they can see the statistics they can see everything.  They can even
have a look outside of hours and see that half their hosts are dead and
yet the statistics show everything is hunky-dory.  So they get very
confused because you haven't told them to ignore the other stuff out of
hours.  But it isn't a lie, according to you, it's "creative accounting." 

>> And I can think of one reason that suppressing this information is a bad
>> idea.  Any trouble outside of those hours may be a sign of an impending
> 
> Nobody said anything about supressing any information.

Yes you did.  This select group of people do not get to see information
about failures that occur outside of hours.  The information that there
have been such failures is *suppressed*.  They do not have the opportunity
to ask "Why do our machines keep failing at night" because they do not
even know it is happening.  It might not directly affect their normal
operations but it is a cause for concern. 

If any of the clients we provide monitoring services to ever found us
pulling a trick like that, where we did 24-hour monitoring but only told
them of faults during their working day, they'd be very upset.  They
might even accuse us of lying to them, because they want to know if
there are any signs of impending failure. 

>> What answer did YOU give him? 
> 
> None.. since the software doesn't actually do what he's asking for,

If Nagios honours work intervals when plotting statistics, then it does.
A single instance of Nagios does not, but two instances will. 

> As to your suggestion that I killfile you... I'm quite happy to simply 
> skip over your posts.

Killfiles require far less effort in the long run. 

> I'm more concerned about the new users who are unaware of your posting 
> habits and might be discouraged from participation by you.

Isn't defining what others may read known as "censorship?"  Which is
pretty much what you tried to do when you told me to stop posting.  And
you're not doing it for yourself but for the good of others who might be
led astray into thinking forbidden thoughts - the excuse of censors
throughout history. 

How about you try using facts and logic to prove what I post is wrong?
Ooops, sorry, that's not one of your strong points.  Your strong point is
snipping a helpful answer of mine then accusing me of not giving a
helpful answer.  Another strong point of yours is pretending that
preventing some data being incorporated into statistics is not suppressing
information - it may be information that somebody has requested to be
suppressed, but it is still suppressed. 

The strange thing about this thread is that in most of your posts you
seem fairly clueful. 

-- 
Paul Allen
Softflare Support 




-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
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