check_postfix_mailqueue

I am starting to use this plugin and I find it extremely useful.
I checked alexlehm's version and I think it's necessary to set a threshold to warn on bounce message count and not warn since first bounced message.
I added a argument (-b):

https://gist.github.com/altmas5/1006fc44c035b6b403099bfeca5f60fb

If an email is delivered between 'find' seeing it and stating it, you get messy output. find has the '-ignore_readdir_race' switch to address this.

@@ -132,10 +132,10 @@
done

# Get values
-deferred=`(test -d deferred && find deferred -type f ) | wc -l`
-active=`(test -d active && find active -type f ) | wc -l`
-corrupt=`(test -d corrupt && find corrupt -type f ) | wc -l`
-hold=`( test -d hold && find hold -type f ) | wc -l`
+deferred=`(test -d deferred && find deferred -ignore_readdir_race -type f ) | wc -l`
+active=`(test -d active && find active -ignore_readdir_race -type f ) | wc -l`
+corrupt=`(test -d corrupt && find corrupt -ignore_readdir_race -type f ) | wc -l`
+hold=`( test -d hold && find hold -ignore_readdir_race -type f ) | wc -l`
bounced=`cat /var/log/maillog | grep bounced | wc -l`
}

I have made an update to alexlehm's version of this script to include a -q parameter with which to select which queue on which to alert.

This change arises from an issue we have seen at my company where the postfix server gets flooded with several thousand messages at once. Postfix continues to process them, so they do not get moved to the deferred queue, but the active and incoming queues grow to almost 2GB apiece.

The new version can be downloaded here: https://gist.github.com/mlwood98/c5cb1e80be595b625c60

Hi,

I think plugin is so good !
But i have problems with permissions (I think).
ls -l /var/spool/postfix returns postfix:root for all folders.
But if I give read permission to root group, script always returns OK and 0 for all. I run script as root user.

Thanks for your return.

To run the plugin from nrpe it is necessary execute it with the sudo permission or the plugin respond always OK with all zero value

Works well, this script checks the mailqueue by directly counting the files in the queue directories, which has the advantage that is will probably be much faster for queues contains 1000s of mails.

Two minor things (not sure if this is right place to report bugs), the command parsing calls the script itself which doesn't make any sense if a parameter is ignored and the check reports 0 for all queues if the directories are not readable due to incorrect permissions or user, in this case /var/spool/postfix is usually readable, but the queue subdirs are not.

A revised script is available here: https://gist.github.com/alexlehm/8084195