I'm trying to configure availability monitoring of three servers, where the severity of the events is based on the number of them which are down - so if one is down then the one event is posted as 'Warning', if two are down then the two events are posted as 'Average', and if all three are down then the three events are posted as 'High' severity. I set this up by creating three triggers per host, then using trigger dependancies to determine which trigger fires - so the Warning event is dependant on the Average event, which is dependant on the High event. Watching the dashboard, it seemed like this was working.
I then created an action which would notify on these events, and had to configure escalations to keep pushing out the event every 10 minutes (due to a requirement that events need to keep sending until cleared). I started to receive event notices, which was good, however I received too many of them, once the severity was increased - shutting down one device gave one notification, shutting down 2 gave 4 events (two at 'Warning' and two at 'Average' severity)
Is this a bug or do I need to configure it differently? Its like the trigger dependancy isn't clearing the event when the dependancy occurs, so the escalation keeps firing - but not sure if thats a bug or by design. This is on a Zabbix 2.0.11 install, running on a RHEL5 server with a PostgreSQL backend db.
I then created an action which would notify on these events, and had to configure escalations to keep pushing out the event every 10 minutes (due to a requirement that events need to keep sending until cleared). I started to receive event notices, which was good, however I received too many of them, once the severity was increased - shutting down one device gave one notification, shutting down 2 gave 4 events (two at 'Warning' and two at 'Average' severity)
Is this a bug or do I need to configure it differently? Its like the trigger dependancy isn't clearing the event when the dependancy occurs, so the escalation keeps firing - but not sure if thats a bug or by design. This is on a Zabbix 2.0.11 install, running on a RHEL5 server with a PostgreSQL backend db.