Hi,
I'm trying to create a trigger based on the % availability of a group (ie cluster) of servers. This is where "availability" is something we define, but readable by Zabbix.
What I've tried so far:
I started by creating a calculated item on each monitored host (member of the host-group) that will be "1" if everything is OK or "0" if not. My next step was to use an aggregated item on the Zabbix server for that hostgroup to average that calculated item.
The problem:
When a host goes down (ie stops communicating), Zabbix either continues to use the last value it received in the aggregate calculation, or stops including the dead host in the averaging, either-way I cannot apparently use this method to work out what % of my servers are available.
I think I could create a calculated item for every host by hand and then logically and that with a zabbix ping check. I could then have calculated item to average that result. That would require me to maintain a large list of calculated items, something I am trying to avoid.
Any ideas?
Tom.
I'm trying to create a trigger based on the % availability of a group (ie cluster) of servers. This is where "availability" is something we define, but readable by Zabbix.
What I've tried so far:
I started by creating a calculated item on each monitored host (member of the host-group) that will be "1" if everything is OK or "0" if not. My next step was to use an aggregated item on the Zabbix server for that hostgroup to average that calculated item.
The problem:
When a host goes down (ie stops communicating), Zabbix either continues to use the last value it received in the aggregate calculation, or stops including the dead host in the averaging, either-way I cannot apparently use this method to work out what % of my servers are available.
I think I could create a calculated item for every host by hand and then logically and that with a zabbix ping check. I could then have calculated item to average that result. That would require me to maintain a large list of calculated items, something I am trying to avoid.
Any ideas?
Tom.