Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo
hello good people,
i have been running nagiosgraph for years, but this issue i dont get, and became stuck.
I have a service called "Number of open files", which has no performance data, only output like so:
"Number of open files: 3540"
"Number of open files: 3540"
I have a corresponding entry in the map file:
/output:Number of open files: (\d+)/
and push @s, [ 'open_files',
[ 'count', GAUGE, $1 ] ];
However, five days ago Nagiosgraph out of the blue one afternoon desided not to update the rrd file no more - remark this is only for this service and another similar one. Output is shown perfectly inside the Nagios gui, and also when run from commandline. I have checked:
) there are write permissions for the rrd file
) there are write permissions for the catalogue of the rrd file
) there are enough diskspace left on device for writing updates
) moved the entry in the map file to the top, just beneath the standard entries
) tried to move the rrd file away, so nagiosgraph could start a fresh, but nothing happened
When enabling insert logging for the particular service via nagiosgraph.conf, all i see in the nagiosgraph.log file is:
Wed Nov 2 22:33:29 2011 insert.pl debug insert.pl processing started
Wed Nov 2 22:33:29 2011 insert.pl debug getrules(/usr/local/etc/nagios/nagiosgraph/map)
Wed Nov 2 22:33:29 2011 insert.pl debug processdata (1 lines)
Wed Nov 2 22:33:29 2011 insert.pl info processed 1 of 1 lines
Wed Nov 2 22:33:29 2011 insert.pl debug insert.pl processing complete
The map entry and the script producing the output have been running without changes for more than a year. No change i am aware of could have triggered this misbehave…
did you update nagios to 3.3.1? processing of performance data and plugin output changed between nagios 3.2.3 and 3.3.1. many people consider the 3.3.1 behavior to be a bug.
thats new to me - yes i have upgraded to 3.3.1 (on freebsd) however, this was a many days/weeks before graphing stopped. And i wonder, how this can be the answer since everything else is as it used to be.
Is there a place where its likely to become clever on this change?
see this thread:
try creating a wrapper around your plugin that captures the plugin output and formats it as standard performance data.
and as of 3.3.1, what is standard performance data supposed to be?
the standard for performance data has not changed. details are in the nagios documentation and in the nagiosgraph README file (see the section 'Adding Service Types'). for your example, it could be as simple as this:
Number of open files: 3540 | openfiles=3540
or, if you wanted to be fancy and include warn, crit, min, and max, it would look like this:
Number of open files: 3540 | openfiles=3540;5000;10000;0;12000
if your service check returns inodes as well as open files, then you might have this:
Number of open files: 3540 | openfiles=3540;5000;10000;0;12000 inodes=3000;4000;8000;0;10000