devmon-support

[Devmon] devmon-0.3.0-beta2 - thresholds bug - resolved

[Dawid M. <daw...@gm...>]

Hi All,

Regarding that thread

http://sourceforge.net/mailarchive/message.php?msg_name=470FBEAF.20906%40cs.toronto.edu

We think that solution may be achieved by only moving  one line of code
just few steps up:

In the file : dm_tests.pm
in function apply_thresh_rep
line number: 2312 - my ($op, $limit) = ($1, $2); should be moved to line
number 2308.

It resolved problem for us :)

Thanks,

Warm regards,
Kasia Michalik
Dawid Mentel

Re: [Devmon] devmon-0.3.0-beta2 - thresholds bug - resolved

[Buchan M. <bg...@st...>]

On Wednesday 12 December 2007 13:38:50 Dawid Mentel wrote:
> Hi All,
>
> Regarding that thread
>
> http://sourceforge.net/mailarchive/message.php?msg_name=470FBEAF.20906%40cs
>.toronto.edu
>
> We think that solution may be achieved by only moving  one line of code
> just few steps up:
>
> In the file : dm_tests.pm
> in function apply_thresh_rep
> line number: 2312 - my ($op, $limit) = ($1, $2); should be moved to line
> number 2308.

xbgmsharp made this change in svn about 2 hours before you sent your mail:

http://devmon.svn.sourceforge.net/viewvc/devmon/trunk/modules/dm_tests.pm?r1=3&r2=10

It looks equivalent.

Regards,
Buchan

[Devmon] New Template F5

[Francois L. <ka...@ka...>]

Hello,

I have commit a new template call f5-bigip-lite and fix f5-bigip
it is usefull with a hobbit client rpm i made for F5.
If people are interest i could publish it.

Why a lite version. When you use the bigip template an a equipement 
which have a lot of pool and vip, you  will get error. So the was to 
made a light snmp test and add all standard stuff  with hobbit client 
since F5 bigip is a linux Redhat3

All of the change have been commit on the svn avalaible at:
http://devmon.svn.sourceforge.net/viewvc/devmon/trunk/templates/

Regards,
Francois

Buchan Milne a écrit :
> On Wednesday 12 December 2007 13:38:50 Dawid Mentel wrote:
>   
>> Hi All,
>>
>> Regarding that thread
>>
>> http://sourceforge.net/mailarchive/message.php?msg_name=470FBEAF.20906%40cs
>> .toronto.edu
>>
>> We think that solution may be achieved by only moving  one line of code
>> just few steps up:
>>
>> In the file : dm_tests.pm
>> in function apply_thresh_rep
>> line number: 2312 - my ($op, $limit) = ($1, $2); should be moved to line
>> number 2308.
>>     
>
> xbgmsharp made this change in svn about 2 hours before you sent your mail:
>
> http://devmon.svn.sourceforge.net/viewvc/devmon/trunk/modules/dm_tests.pm?r1=3&r2=10
>
> It looks equivalent.
>
> Regards,
> Buchan
>
> -------------------------------------------------------------------------
> SF.Net email is sponsored by:
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services
> for just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
>
>   

Re: [Devmon] New Template F5

[Francois L. <ka...@ka...>]

Here are my cisco pix template ciscopix6 cisco-pix702 cisco-pix706

I have about 250 pix to monitor and i really need a lot of information 
and even more graph to get history on each device.

Re: [Devmon] New Template cisco-pix6 cisco-pix702 cisco-pix706

[xbgmsharp <xbg...@gm...>]

I figure out how to graph out of the box the memory.

First hobbit look for this word:
Physical or Swap or Actual
then i take the pourcentage of it. only digit (eg 25%).

So in your memory message, you need to have line like this:

Physical {nice_mem_used_per}%

So if you don't want to show it do like for CPU
<!-- DEVMON to RRD Physical {mem_used_per}%
-->

In your tranfrom:
mem_used_per : MATH : ({mem_used} x 100) / {mem_total} : 0


EG the documentation:
    Decimal precision can also be controlled via an additional
    variable seperated from the main expression via a colon:

       transTime : MATH : ((({sysUpTime}/100) ^ 2 ) x 15) + 10 : 4
 
    This would ensure that the transTime alias would have a
    precision value (zero padded, if needed) of exactly 4 characters
    (i.e. 300549.3420).  The default value is 2 precision characters.
    To remove the decimal characters alltogether, specify a value of 0.

And a graph appear.



FOr the CPU like in many template you need:

<!-- DEVMON to RRD <br>CPU 5 min average: {CPUTotal5Min}
-->

It will graph %  as load.


Regards
Francois


Francois Lacroix a écrit :
>
>
> Here are my cisco pix template ciscopix6 cisco-pix702 cisco-pix706
>
> I have about 250 pix to monitor and i really need a lot of information 
> and even more graph to get history on each device.
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> SF.Net email is sponsored by:
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services
> for just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> ------------------------------------------------------------------------
>
> _______________________________________________
> Devmon-support mailing list
> Dev...@li...
> https://lists.sourceforge.net/lists/listinfo/devmon-support
>   

Re: [Devmon] Active connexion questions

[Buchan M. <bg...@st...>]

On Friday 14 December 2007 19:22:50 xbgmsharp wrote:
> Hello all,
>
> A lot of work have been done today. easy graphing CPU load% and memory %.
>
> But here is a another point.
> It is usefull do know (graph) the number of active connexion on any
> network device.
> In the mean time i am trying to do it for Pix and F5.
>
> The problem is:
>
> Hobbit know how to graph process count, it is done via the procs/ports
> tests (colunm) then by sending data intead of status.
> What do you think should be done:
>  1. make devmon sending directly a data to graph, so no test colunm only
> graph
>  2. Create a test (colunm) and then do nvc or extra-scritp in order to
> graph.


In my pix/fwsm template (I posted them), I am using 'conns' name, with Devmon 
RRD (requiring an extra-script such as my extra-rrd.pl) to graph the number 
of connections.

Since I don't think it is currently feasible to get all graphs we want into 
Hobbit with existing hobbit support, I think it would be best (for now) to 
ship a devmon extra script (one which uses the RRD information created by the 
rrd option). In future, we can integrate the support for this into Hobbit, or 
switch to the new trends feature in 4.3.

I don't think we should abuse the ports test name for this, because the data 
is different (just the number of connections, no detail etc.).

If you want to alert on the same data, I don't think sending a 'data' message 
will work. So, 'status' is the only option for now.

I'd rather ship a release that has graphs that work as soon as possible (even 
if that requires running one extra script in hobbit) than delay shipping a 
solution because we want a perfect one. Im-perfect graphs are better than 
none.

Regards,
Buchan

Re: [Devmon] Active connexion questions

[Simeon B. <sbe...@mc...>]

Buchan Milne wrote:
> On Friday 14 December 2007 19:22:50 xbgmsharp wrote:
>> Hello all,
>>
>> A lot of work have been done today. easy graphing CPU load% and memory %.
>>
>> But here is a another point.
>> It is usefull do know (graph) the number of active connexion on any
>> network device.
>> In the mean time i am trying to do it for Pix and F5.
>>
>> The problem is:
>>
>> Hobbit know how to graph process count, it is done via the procs/ports
>> tests (colunm) then by sending data intead of status.
>> What do you think should be done:
>>  1. make devmon sending directly a data to graph, so no test colunm only
>> graph
>>  2. Create a test (colunm) and then do nvc or extra-scritp in order to
>> graph.
> 
> 
> In my pix/fwsm template (I posted them), I am using 'conns' name, with Devmon 
> RRD (requiring an extra-script such as my extra-rrd.pl) to graph the number 
> of connections.
> 
> Since I don't think it is currently feasible to get all graphs we want into 
> Hobbit with existing hobbit support, I think it would be best (for now) to 
> ship a devmon extra script (one which uses the RRD information created by the 
> rrd option). In future, we can integrate the support for this into Hobbit, or 
> switch to the new trends feature in 4.3.
> 
> I don't think we should abuse the ports test name for this, because the data 
> is different (just the number of connections, no detail etc.).
> 
> If you want to alert on the same data, I don't think sending a 'data' message 
> will work. So, 'status' is the only option for now.
> 
> I'd rather ship a release that has graphs that work as soon as possible (even 
> if that requires running one extra script in hobbit) than delay shipping a 
> solution because we want a perfect one. Im-perfect graphs are better than 
> none.
> 

We've had good luck with using NCV to add graphs. The pix template that I've 
recently submitted is designed so that the message is formatted with that in 
mind. The downside is that it does require you to edit hobbitgraph.cfg and 
hobbitserver.cfg in order to add the NCV configuration options. On the positive 
side, once you've worked out a couple of these it's not too hard to remember how.

-- 
S i m e o n  B e r k l e y

Systems Engineer
McClatchy Interactive
phone:  919-861-1244
fax:    919-861-1300
mobile: 919-302-3063
e-mail: sbe...@mc...
AIM:    sberkleymi
www.mcclatchyinteractive.com

Re: [Devmon] Active connexion questions

[Simeon B. <sbe...@mc...>]

xbgmsharp wrote:
> It sound to me like a nice idea.
> Could you share you script.
> Where is your cisco pix template?
> I have probleme pix 7.0.2, do you?
> 

Attached is the pix template that I'm using, which is designed around v6.3. I 
don't have access to a pix 7.x. Here is how we get values into rrd's using our 
template:

If you look at the fwmem, fwcpu, and fwconn message files, they are formatted to 
send the status message formatted so that NCV will pick up the values. eg:

  mp: {MemPercUsed}
  mt: {MemTotal}

To get this to populate rrd's you need an NCV line and related configuration in 
hobbitserver.cfg:

  TEST2RRD="cpu=la,(...bunch of stuff...),fwconn=ncv,fwcpu=ncv,fwmem=ncv"
  GRAPHS="la,(...bunch of stuff...),fwconn,fwcpu,fwmem"
...
  NCV_fwcpu="c1:GAUGE,c5:GAUGE,c15:GAUGE"
  NCV_fwconn="c:GAUGE,h:GAUGE"
  NCV_fwmem="mt:GAUGE,mp:GAUGE"

Note how the rrd definitions have matching datasource names (mp,mt,c1,c5,etc)

To get hobbit to generate graphs with these rrd's, you need appropriate graph 
definitions in hobbitgraph.cfg:

  [fwcpu]
    TITLE Cpu usage
    YAXIS Percent
    -u 100
    -l 0
    DEF:c1=fwcpu.rrd:c1:AVERAGE
    DEF:c5=fwcpu.rrd:c5:AVERAGE
    DEF:c15=fwcpu.rrd:c15:AVERAGE
    LINE1:c1#F0000F:Cpu usage
    GPRINT:c1:AVERAGE:mean\: %%%.0lf
    GPRINT:c1:LAST:current\: %%%.0lf
    GPRINT:c1:MAX:peak\: %%%.0lf\l
  [fwconn]
    TITLE Connections
    YAXIS Connections
    DEF:c=fwconn.rrd:c:AVERAGE
    DEF:h=fwconn.rrd:h:AVERAGE
    CDEF:cK=c,1000,/
    CDEF:hK=h,1000,/
    AREA:c#0055FF:Connections
    LINE1:c#000000
    GPRINT:cK:AVERAGE:mean\: %.3lfK
    GPRINT:cK:LAST:current\: %.3lfK
    GPRINT:cK:MAX:peak\: %.3lfK\l
    LINE2:h#FF0000:Historic Peak\:
    GPRINT:hK:MAX: %.3lfK\l
  [fwmem]
    TITLE Memory usage
    YAXIS Percent
    -u 100
    -l 0
    DEF:mp=fwmem.rrd:mp:AVERAGE
    DEF:mt=fwmem.rrd:mt:AVERAGE
    CDEF:mtM=mt,1048576,/
    AREA:mp#0055FF:Used
    LINE1:mp#000000
    GPRINT:mp:AVERAGE:mean\: %%%.0lf
    GPRINT:mp:LAST:current\: %%%.0lf
    GPRINT:mp:MAX:peak\: %%%.0lf\l
    GPRINT:mtM:LAST:Total Memory\: %.0lfM\l

I prefer the NCV approach because it is clearly defined and is a stable part of 
hobbit. The downside is having to maintain configuration for one thing in two 
config files. Our decision to stick with NCV was based upon keeping the 
complexity in the configuration instead of the executables so we don't have to 
track changes in a bunch of 3rd party scripts. Devmon has been an exception to 
the 3rd-party-scripts rule since it handles a major system (snmp) that isn't 
(yet) a core part of hobbit.


-- 
S i m e o n  B e r k l e y

Systems Engineer
Mcclatchy Interactive

Re: [Devmon] Active connexion questions

[xbgmsharp <xbg...@gm...>]

Hello all,

A lot of work have been done today. easy graphing CPU load% and memory %.

But here is a another point.
It is usefull do know (graph) the number of active connexion on any 
network device.
In the mean time i am trying to do it for Pix and F5.

The problem is:

Hobbit know how to graph process count, it is done via the procs/ports 
tests (colunm) then by sending data intead of status.
What do you think should be done:
 1. make devmon sending directly a data to graph, so no test colunm only 
graph
 2. Create a test (colunm) and then do nvc or extra-scritp in order to 
graph.

Regards,
Francois