Re: Monit and SQL query

[Jan-Henrik Haukeland]

Yes, more or less. To have Monit send an alert message that is printed from 
your script, follow this receipt 

1) Add a custom mail format to the check, where the message body is just 
$DESCRIPTION:

 check program foo with path "/path/to/foo.sh"
     alert address@hidden mail-format { message: $DESCRIPTION }
    if status != 0 then alert

2) Make your script write the error message you want to send to stderr. Here is 
a demo script I used to test this:

foo.sh:

 #!/bin/bash

 echo 1>&2 "Hello this is an error message from $0";
 exit 1;

3) The alert mail you get from Monit will look something like this.

 status failed (1) for /path/to/foo.sh -- Error: Hello this is an error message 
from /path/to/foo.sh

The prefix 'status failed (1) for /path/to/foo.sh -- Error: ' is hardcoded in 
validate.c:check_program and if it bothers you, just remove the prefix. 

That is replace this line

   Event_post(s, Event_Status, STATE_FAILED, status->action, "status failed 
(%d) for %s -- Error: %s..", s->program->exitStatus, s->path, buf);

With this one:

  Event_post(s, Event_Status, STATE_FAILED, status->action, "%s", buf);

and recompile Monit. 

Hope this helps.


On Sep 16, 2012, at 4:33 PM, Charles Pino <address@hidden> wrote:

> i had it somehow working, but the alerts message / email is somewhat not 
> really nice to look at.
> 
> 
> Status failed Service foo 
> 
> Date: Sun, 16 Sep 2012 21:12:38
> Action: alert
> Host: orecx
> Description: status failed (1) for /root/mysql.py. Error: 1
> .
> 
> Your faithful employee,
> Monit
> 
> 
> my question is. can i change the status failed service to my own message? as 
> well as 
> Description: status failed (1) for /root/mysql.py. Error: 1 --  id like to 
> have it say that the db hasnt been any writes in the past 12  hrs etc etc.