Auto recover an PROBLEM state for an action

[bignose]

I'm doing log monitoring like this:

Swatch runs on my servers and monitors my logs. When a certain string is found in the logs we run zabbix_sender.

This means that the trigger that watches for differences effectively stays in the PROBLEM state because it would need the same data twice in a row in order to recover.

Looking for some suggestions..

[bignose]

maybe some kind of trick with hysterisis and nodata ?

[bignose]

hrmm, this seemed to work but i'm unsure of how great of a plan it is...

Add this to cron on the machine...

zabbix_sender -z 10.33.33.52 -s java3.toronto.terida.net -k swatch_log_monitor -o " -- " so it sends it every minute. if its interuped by swatch doing its thing, the trigger gets fired then it gets set back to normal.