Reporting and misc rave.

[John P. Rouillard]

In message <D00B29FE26656E4783F128B118D8D43703B3AF6E at acexp001.portfolio.base>,
Stanley.Hopcroft writes:
>I am writing with mainly a rant about Graphing and Reporting.
>
><Mainly OT rant>

OT -> on topic in this case IMHO 8-).

>1 About graphing with Nagios
>Why would one bother when 
>  1.1 Cacti does such a good job

Well cacti does a good job but I hate the thought of having to poll
using both cacti and nagios:

  it doubles the load on the network and the target system
  requires twice the setup and maintenance

I agree for snmp polling especially cacti is much better.
Autodiscovery of disks, network interfaces etc and all of that is very
nice.

I assume your solution for graphing is to use cacti to poll all the
devices for which you want to keep performance data (which in my case
would be everything 8-) so I can use holt-winters analysis).

>  1.2 Nagios could check the Cacti RRDs with either check_rrd, or by an 
>  outboard (Cron scheduled) RRD poller that submits passive service
>  check results

I had problems trying to get check_rrd to work. I was having problems
finding the perl modules it needed IIRC.

>  1.3 the graphs can be associated with Nag service checks by either
>
>   - explict URL of the Cacti graph in the service check output
>
>   - for the adventurous, a Wiki front end that displays some of the 
>     Nag CGI service status and a link to the Cacti graph. 

You forgot a couple: use the "Extra Service Actions" link on the
Service Information page to link to cacti for the host using a link
like:

  http://cacti.my.server/cacti/graph_view.php?action=preview&filter=monitored_host

A second way is available due to a patch I got added to nagios.  If the
page footer and header are executable, they are executed (as a cgi)
when the page loads and can be used to load graphs/menus
etc. automatically based on service, host, user or other CGI
variable.

As a demo I used drraw to create nice dashboard and plot different data
items from separate rrd's (e.g. a graph of all ping times to all hosts
at a single site and a second network throughput graph displayed nice
and large together on a single page) for easy analysis. It was displayed
anytime I went to the PingCheck page for a couple of hosts.

>  As a footnote, since Cacti supports RRD 1.2 with built-in supported

Hmm, must have missed that last I knew it only supported the 1.1
series.  Cool thanks for the update.

>  Holt Winters forecasting RRAs, the poller could be smart and simply
>  check the exception Data store to see if the current rate is in fact
>  outside the normal seasonal variation (computed by the Holt Winters
>  algorithm inside the RRD).

Yup my idea exactly. Use nagios to monitor for worst case (e.g. ping
time > 1 second) and put the nagios data into rrd's to use the
Holt-Winters forecasting to look for early signs of a problem before
they trigger the nagios thresholds.

>2 About reporting
>
>After writing a lot of code in Nagios::Report to extract and report on
>Nagios availability data it occurs to me that a better way of doing
>Reporting is to
>
>  2.1 put the availability data in a DB table (prob with an
>auto-incremented index)

What would be nice is a sqlite <http://www.sqlite.org/> datastore and
event broker. Low maintenance requirements, no server to configure, has
most of SQL 92 implemented.

>  2.2 use either
>    2.2.1 ad-hoc SQL queries, or
>    2.2.2 the reporting package of your choice (eg iReport)
>
>I hope that Nagios::Report will be enhance to take advantage of
>DBD::Ram, a Perl module that very easily gets a CSV file with LWP and
>sticks it in an in core DB that can almost as easily be used as a
>Data source to insert rows into the DB of your choice (MySQL, or
>whatever).

Neat idea. Could use: DBD-SQLite-1.11 as the back end too. Sadly I
don't see much for reporting tools for SQLite. However it does have
ODBC and JDBC connectors and a bunch of bindings for tcl/tk, perl,
python, ruby etc. SQLite is a nice "throw away" database easy to set
up, load, query and destroy.

></Mainly OT rant>

Not off topic as far as I am concerned. It all deals with effective use
of nagios to monitor the network.

				-- rouilj
John Rouillard
===========================================================================
My employers don't acknowledge my existence much less my opinions.


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
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