IT Services

[bobrivers]

Hi,

I'm trying to use IT Services. I created the following structure:

BR
...Servers
......ServerA
.........TriggerA
.........TriggerB
.........TriggerC

It's working, but, the way zabbix calculates the SLA seems to be wrong. At least I don't agree with that...

In the example above, the ServerA SLA will be equal to the worst child SLA... If triggerA is 100%, triggerB is 100% and triggerC is 80%, the ServerA SLA will be 80%...

IMHO, the ServerA SLA should be 96,66% (sum of all triggers divided by the quantity). I'm assuming that all trigger are equally important.

I tried to change the calculation algorithm (if at least, if all, etc), but the results are basically the same (using "if all" the ServerA will remain with 100%).

Am I missing something?

TIA,

Bob

[bobrivers]

ping

Just to give the message a new chance...

Nobody?

TIA,

Bob

[Aly]

ServerA is dependent on triggers, if one of the triggers is down(problem) - we assume ServerA is down too. So SLA goes down for this service.

[bobrivers]

Hi, thanks. No I know that it's working how it should work.

Just trying to help to improve zabbix: I understand your point. It makes sense, but I don't agree with it. At least I think that it should be more flexible.

Ok, if I have services that are directly dependent on triggers, it's reasonable to think that if one of then is down, the entire service is down.

But, if we have nested services, IMHO, zabbix should calculate the sla and weigh up all the individuals sla. I think that it's more fair.

Right now the tree view structure it's only suitable to organize individual services, not to measure a group sla.