Ideas on how to monitor backups

[hertell]

Hi all!

I'm handing a couple of linux-clients rsync-backups with a perl-script. When succesfully finished, it posts me the size of each successful backup with zabbix_sender. This part works great.

The problem (i think) i have is that if the backup fails, eg. the rsync-process get's screwed up in any way, or the destination hdd is ful or the backup just did not start for any reason, i won't get any notice to zabbix.

Is there any way to start a count-down-timer in zabbix that when it reaches Zero (or > 24h) then it would fire a trigger. I was thinking of having zabbix_sender reset this timer each time the backup was succesful, and if it would fail in any way, it would just not do this, and when the calculated bacup-time has reaced it's end, the trigger would get fired.

I was first thinking of having a cron-job that would monitor a file with a timestamp of a succesful backup. if this timestamp would excceed eg. 24, then it would trigger an alarm..

But this method just feels a bit clumsy..

Any new/modified ideas are welcome :-)

[alixen]

Hi,

You can use nodata() function to check it.
Your trigger would be something like:
{host:item.nodata(86400)}#0

86400 seconds is 24 hours. You may use a slightly larger value to avoid false alerts.

Check the manual for details.

Regards,
Alixen

[hertell]

Hi,

That was a great tip! I'll dig into that :-)