Feature #165

Audit facility: research

Added by Charles Atkinson almost 8 years ago. Updated about 1 year ago.

In Progress
Start date:
Due date:
% Done:




#1 Updated by Charles Atkinson almost 8 years ago

  • Subject changed from nagios plugin: create to nagios plugin: research

Is it possible to specify a bung audit utility which can both run independently and (maybe) as a back-end for a nagios plug-in?

#2 Updated by Charles Atkinson about 6 years ago

  • Subject changed from nagios plugin: research to Audit facility: research

#3 Updated by Charles Atkinson about 2 years ago

Charles and Daniel discussed. Daniel would like a backup audit facility.

#4 Updated by Charles Atkinson about 1 year ago

Subject:     Re: Zabbix Agent User Parameters
Date:     Thu, 3 Sep 2020 12:13:54 +0530
From:     Daniel Estis <>
To:     Aurinoco Systems <>

I meant to solve the issue of backup task has not been executed for whatever reason. And hence no email being sent. We spoke about this.... I am working now on a new version of backup script coded in Powershell and I am thinking about an SFTP box that will accumulate reports from all hosts. Then another script will analyse these reports and send a concentration once a while. What do you think?


On Thu, Sep 3, 2020 at 9:35 AM Aurinoco Systems < <>> wrote:

    Assuming you mean bung backups ...

    We could but bung already sends emails (SUCCESS or otherwise) and
    logcheck emails bung's warning and error messages written to
    syslog.  Would another notification add value?


    Aurinoco Systems
    Auroville Foundation Bhavan
    Auroville, Tamil Nadu 605101
    +91 413 2000015

    On 03/09/2020 09:28, Daniel Estis wrote:
>     Hello Charles,
>     Just a very crude idea that came to my mind this morning... Can we
>     use Zabbix Agent User Parameters to deliver backup results?
>     best,
>     Daniel

#5 Updated by Charles Atkinson about 1 year ago

  • Assignee changed from Charles Atkinson to Daniel Estis

It is a great idea now, as it was six years ago!

I understand you are writing of Windows. I am replying for Linux. The principles are the same

The audit facility should also audit database (mariadb, postgres ...) and OpenLDAP installed and backups not configured.

Rather than having a central location to receive reports (your SFTP box) I would
  • Create a script to be run by Zabbix on the monitored computer to
    • Audit what has been installed (mariadb, OpenLDAP, postgres ...) and, for each, report
      • How it is being backed up (the backup conf)
      • Status of the last backup (time run, OK/WARN/ERROR)
    • On Dom0s only, similar for self backup and each DomU backup

Implementing this via Zabbix would keep everything in the same tool regards logging and notifications

#6 Updated by Daniel Estis about 1 year ago

  • Assignee changed from Daniel Estis to Charles Atkinson

I would also prefer to keep all in the same tool. But I will need to learn how to communicate a parameter to Zabbix. Maybe, if you do progress on this, we can meet and you briefly explain me how Zabbix works and what might be required from a backup script to communicate data to Zabbix.

#7 Updated by Charles Atkinson about 1 year ago

  • Status changed from New to In Progress
  • Assignee changed from Charles Atkinson to Daniel Estis
As discussed, Zabbix can be extended to run custom commands on monitored hosts. Aurinoco has done so to check postfix queues.
  • On the Zabbix server, template "Template App Postfix" has items
    • Postfix active queue
    • Postfix bounced mails
    • Postfix deferred queue
    • Postfix sent mails
  • Taking "Postfix deferred queue" as an example, the item is populated by getting postfix.deferred from the Zabbix Agent running on the monitored host
  • On the monitored host /etc/zabbix/zabbix_agentd.d/postfix.conf includes
    UserParameter=postfix.deferred[*],/usr/sbin/postqueue -p | egrep -c "^[0-9A-F]{10}[^*]" 
    That defines the value of as the output of the postqueue and egrep commands
  • Template "Template App Postfix" also has a trigger "Postfix: Too many deferred mails on {HOST.NAME}" which fires when the average number of deferred mails for the last three hours exceeds two (in retrospect not the best choice!)
The value of is an deferred. For backup monitoring we would probably want two values
  • An integer representing OK, WARN or ERROR and displayed as such
  • Text detailing the WARN or ERROR condition

The custom command can be a script so getting the values is completely flexible

#8 Updated by Daniel Estis about 1 year ago

One more value. Date of last backup execution.

Also available in: Atom PDF