From: Shane M. <sh...@mc...> - 2017-09-22 06:56:39
|
On 21/09/17 11:04, Roy Sigurd Karlsbakk wrote: > I'm looking at 'services' and it seems to be somewhat what I'm > looking for, but I don't understand how I can aggregate a bunch of > tests for a varying bunch of hosts and their services. We currently > have around 300 servers, somewhat 60/40 linux/windows, and new ones > pop up as old ones no longer in use are discarded quite regularly. I think what you want to do is create Host Groups[1] that group your servers however you want (By location, operating system, whatever). You can then create Aggregate checks[2] based on the host group and a common item that all the hosts in the group have. For example: grpmax["Linux Servers", "system.cpu.load[,avg1]", last] That will give you a single value: the maximum CPU load from all your servers. > So I guess what I'm looking for, is just something like "please give > me a list of hosts in hostgroups x,y with severity > n" and then feed > that to a service (formerly IT service). I've worked with zabbix for > a while now, but mostly for testing and with a small initial setup of > 65ish servers and 10ish switches/upses etc. Once you have an aggregate value from your host group you can then create a trigger on that value: {hostname:aggregate_name.last()} > 5 The "hostname" can be any host in your host group (I think! I'm doing this from memory!) and "aggregate_name" is the name of the aggregate you created above. Once you create the trigger it will activate if any host in the host group goes over a CPU load of 5. Hope this helps! Shane. [1] https://www.zabbix.com/documentation/3.4/manual/config/hosts/host#configuring_a_host_group [2] https://www.zabbix.com/documentation/3.4/manual/config/items/itemtypes/aggregate [3] https://www.zabbix.com/documentation/3.4/manual/config/triggers/expression |