Condense multiple items into a single trigger

[EtherTrogg]

Is there a way to condense multiple items into one trigger so that I don't have to fill up my trigger page with repetitive entries?

In my case, I'm monitoring the power supply status on an MRV MCC288 layer-1 switch. The MCC288 has four (4) power supplies installed, which I am monitoring using SNMP. Each supply has its own unique SNMP OID. I'd like to monitor each supply, but have one trigger entry that will fire if any of the items fails.

I've tried creating four triggers with the same name: "Power Supply Status". Each trigger is monitoring a different item. This shows only one trigger on the trigger page (which is what I want), but it only "fires" for one item.

Can this be done?

[bbrendon]

Sounds do-able.

It sounds like you'll have 4 items in zabbix. One for each power supply.

One trigger for all power supplies.

You can have multiple requirements or dependencies in a trigger by using "&" or "|"

Here is an example of one I use to use with my pinger item:
({server2:system.cpu.util[].nodata(755)}=1)&({server1inger[cl.verio].last(0)}=1)

Notice the "&". Just list your PS's and separate with "|"

[EtherTrogg]

Sounds do-able.

It sounds like you'll have 4 items in zabbix. One for each power supply.

One trigger for all power supplies.

You can have multiple requirements or dependencies in a trigger by using "&" or "|"

Here is an example of one I use to use with my pinger item:
({server2:system.cpu.util[].nodata(755)}=1)&({server1inger[cl.verio].last(0)}=1)

Notice the "&". Just list your PS's and separate with "|"

Excellent! I'll give this a try. Thanks!

[tighep]

In the case of filesystem space, is there a way to pass the filesystem name into the action? If I consolidate my triggers I can't seem to see which filesystem has filled up.

[otheus]

I have an issue with multiple items in a trigger: When the trigger condition is met, you can a trigger notice for each host in the trigger. That doesn't seem like what I want.

For instance, I have 5 servers that operate in parallel. I want to know if one of the servers' load is 10% above the average of the cluster. So I create an "aggregate item" that exists in a special "aggregate host". Then for each host, I create a trigger that compares the current server to the aggregate item, like

{server1.cpu.load[,avg].last(0)} > 
    {aggregate_server.grpavg[...blahblah...].last(0)}

When the CPU load goes over the average, I should get a trigger for "server1". But I get a trigger ALSO for "aggregate_server", which is undesirable.

Any hack around this?