Monitoring the monitor?

[zervin]

Recently ran into the file size limit exceeded error on the cacti log. I am trying to find the best way to monitor the monitor. The system died on friday afternoon, and I didnt know it wasn't plotting until my mgr pointed it out lol.

What do you guys this is the best way to keep tabs on cacti, and assure everything is running as it should? I have considered logging to syslog, or maybe monitoring the cacti.log.

Does anyone have any suggests or examples of how they keep tabs on cacti itself?

[Linegod]

First, ensure you are using logrotate to rotate our the Cacti log ( http://docs.cacti.net/manual:087:4_help ... _cacti.log ).

Personally, I use Nagios to check the cacti.log for SYSTEM STATS output within a specific timeframe, and checks Time, Hosts, Data Sources, etc for large increases or drops.

The same thing could be accomplished without Nagios, by creating a simple cronjob that checks the last write time on the log, as well as one that parses the output (if you want...)

[zervin]

Yeah what I was planning to do is just run a cronjob to check for any weird errors in the cacti.log.

If there a table somewhere that show all the different preset errors cacti can generate in the log, something I could parse on.

[schurzi]

we check the health of our cacti system with nagios and a script, which does a find on our rrd directory, to find any files with a modification time older then 10 minutes. This should catch almost any possible error in cacti or single datasource.

[Linegod]

Take a look at the 'View Cacti Log File' under 'System Utilities'. Message Type: has the fields.

You may want to look at the code for utilities.php to shortcut your scripting.