Xymon Mailing List Archive search

SPLITNCV and graphs

list Ralph Mitchell
Thu, 16 Apr 2015 23:25:17 -0400
Message-Id: <CAAEjoCV-ZNWpMJLK3=-09dZ6otkOuzufv2PaA=user-1405207e781c@xymon.invalid>

You probably need to create the RRDs manually. I don't recall the details,
but you can specify how often samples should arrive and how many can be
missed before it records not-a-number (NaN).  The default sample interval
is 5 minutes, and I think if you miss two that counts as NaN.  That fits
with your observed loss of graphs.

Xymon doesn't care who created the RRD, so you should be able to create one
with the rrdtool command line program that would work with 15 minute
samples.

Ralph Mitchell


On Thu, Apr 16, 2015 at 10:40 PM, Galen Johnson <user-fc632e705d24@xymon.invalid> wrote:
Thanks for the replies.  Adam's reply regarding the way RRD handles the
data makes sense.  Once I move the test to run every 5 minutes, it did
begin to populate the graph.  I let it run for hours with the 15 minute run
without success (looks like RRD stops caring after 10 minutes).  I'll look
more into RRD to see if there is some easy way to have this work since it's
a valid need to have tests that run less often but you still want to graph
for trending purposes.  Likely just a user education issue.

I looked at the data message but the man page seemed to indicate that this
would only populate a graph only for the trends page.  I may revisit it
since when I stopped trying to use it, I was still running at 15 minute
intervals.

per WJM's comment, It's a GUAGE, actually.  I may look at the devmon
option.  I may not have been clear on the request to have the data be
restricted somehow for the status. It's too easy to have NCV trigger since
'=' and ':' are fairly common separators when one provides output.  I love
the flexibility of NCV but that little foible is a bit annoying.  Also
note, even when I defined NONE, it still creates an rrd file (I used
SPLITNCV).  It would also be difficult to set that up for a string that
looks like this:

Thu Apr 16 22:33:41 EDT 2015

and creates a rrd file named
test,Thu_Apr_16_22.rrd

that is going to create a new rrd file every hour.

Having to filter the content to change it to use the html entities is
(IMHO) a hack that shouldn't be needed.  The devmon test apparently already
requires this so it really doesn't seem too unreasonable for NCV to handle
it the same way.

Just throwing it out there...it's graphing along and I'm happy I was able
to figure it out.

On Thu, Apr 16, 2015 at 2:58 AM, W.J.M. Nelis <user-6956df205d63@xymon.invalid> wrote:
 Hello,

 I'm trying to add a new graph to a test using NCV (SPLITNCV actually).
I read through the graphing tips, I read several emails that provided hints
(not my exact problem but nudged me the right way)...

 long story short.   I finally got he graph to show up on my test
page...but the rrd file doesn't seem to be populating the fields as
expected when I run "rrdtool dump".  The only thing I can figure is that
the test I'm performing doesn't run every 5 minutes but every 15 minutes.
Does that matter?  I would have expected in the worst case that only every
3rd row would have a value.

 What am I missing here?  Note, I dropped my test to run every 5 minutes
to see if this theory is true.

Is the DS type DERIVE or COUNTER? In that case two successive samples,
which are 5 minutes apart, are needed to create one data point in the graph.


 As a side note, it would be extremely helpful if the NCV component
didn't look at the entire status but required a comment block (for example)
to contain it.  That way you could keep your data stream isolated from your
message.  For example:

 <!-- NCV data
metric1: 123
metric2: 456
etc
-->

Perhaps you can use the devmon format, which encapsulates the data to
trend in an HTML comment section. In that case, one would have to specify
"test=devmon" in stead of "test=ncv" in variable TEST2RRD. There is no need
for an NCV_Something variable, as the DS definition is included with the
data to trend.


 My test kept creating an rrd file for a date string that was part of
the status message.  That's a bit annoying.

  There are (at least) two ways to circumvent that problem:
 A) Add "*:NONE" to the NCV_Something variable;
 B) Replace all occurrences of ":" and "=" outside the NCV part by their
HTML equivalent, thus "&#58;" and "&#61;".

Regards,
  Wim Nelis.


*The NLR disclaimer is valid for NLR e-mail messages.*

This message is only meant for providing information. Nothing in this
e-mail message amounts to a contractual or legal commitment on the part of
the sender.

This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. Sender accepts
no liability for damage of any kind resulting from the risks inherent in
the electronic transmission of messages.