n2rrd Plugins

                                

        
        

            
            

                

    

    
n2rrd_plugins


we're running a Nagios installation where the performance data gets passed to n2rrd to be stored in RRD files. n2rrd is able to generate templates etc. etc. on its own, but that behavior might not always be what you want. This set of plugins allows Nagios to monitor the activity of n2rrd for some problems that we experienced.


check_ds_of_rrd


Usage


check_ds_of_rrd [ -v | -d | -h ]


Checks the Round Robin Databases stored in RRD_DIR against the structure templates in RRA_DIR
and complains if the list of data sets (DS) stored in the RRD does not match the DSes in the template.

(n2rrd will automatically introduce extra DSes in the RRDs if the performance data from the plugin, after processing by the PERL code snippets in dirname $RRA_DIR/code, if any, present extra DSes.)

Extra DSes in your RRDs are not a big problem, but they do make the RRD files unnecessarily (?) bigger, and if it's just a typo - say, your check_disk reports '/home', which gets autochanged to '_home', and your dirname $RRA_DIR/graph template expects 'home' -, it can
be quite a puzzle why your graphs remain blank ...


check_dyn_rrd


Usage


check_dyn_rrd [ -w < num > ] [ -c < num > ]


Counts the number of dynamically created templates, both for RRAs and graphs, and raises an appropriate alert if more than the given limits are found.

The limits default to -w 15 -c 30.
n2rrd's feature to create template files automatically is nice to have in some instances, but there also are many situations where you do not want it to deviate from what you set up manually.

This plugin can be used to keep tabs on the total number of dynamic templates and alert you if it grows unexpectedly.


check_stale_rrd


Usage


check_stale_rrd [ -w < minutes > ] [ -c < minutes > ]


Checks the RRD files and raises an alert if some have not been updated for more than the specified limit.
The default limits are -w 130 -c 250.
Note that this is prone to produce duplicate alerts as soon as some service check fails in such a way that it stops producing performance data - use with caution.^