annoying false positives...

[Plexus]

Hi folks,

I am monitoring several cronjobs by sending a "0" string via zabbix_sender to a trapper item. Some of those cronjobs, running non-cyclic at specific times, do have a trigger checking for feedback to corresponding moments and this trigger produces sporadic false positives.

A trigger for a cronjob started at 02:05 and needs 10 minutes maximum for execution looks like this:

({TRIGGER.VALUE}=0 & {my.cronserver.de:my.cronjob.php.nodata(10m)}=1 & {my.cronserver.de:my.cronjob.php.time(0)}=021500) | ({TRIGGER.VALUE}=1 & {my.cronserver.de:my.cronjob.php.nodata(2m)}=1)

In the lastest 500 values of latest data I see that the job´s feedback arrived:

2014.Jul.14 02:06:08	0
2014.Jul.13 02:05:50	0
2014.Jul.12 02:05:55	0
2014.Jul.11 02:05:49	0

Nevertheless this trigger is activated and I do not see the reason why... Thanks in advance for your input.

greetz,
Plex

[Plexus]

In the meantime I found out, that system load by the hostmachine´s backup is causing this error. As I am working with a cluster I will move the active clusternode now by cronjobs according to the backup times.

edit:
this helped a bit but wasn´t the final solution...