Trigger dependecies?

[schneck]

Hi,

i have several simple-check items & triggers for a template, say:

PING {HOSTNAME} unreachable: {TMPL-Std:icmppingsec.max(375)}=0.0
SMTP {HOSTNAME} unreachable: {TMPL-Std:smtp_perf.max(615)}=0
SSH {HOSTNAME} unreachable: {TMPL-Std:ssh_perf,22.max(615)}=0
...

Now, if a host goes down, all those triggers fire and I get several alerts.

What I want is to only get service (eg. SMTP) alerts if the host is up (PING trigger is OFF)

I tried two different approaches, but didn't get anywhre yet:

Using trigger dependencies

This doesn't seem to work because trigger dependecies can only be defined on other triggers which are ON.

Using action conditions

This doesn't seem to work because action conditions can only
* use fully specified triggers, not generic template PING {HOSTNAME} unreachable triggers
* use trigger values of the current trigger, not some-other-trigger = OFF

Possible solutions would be:

* don't use templates and define every trigger/action for every host
* define an additional trigger PING {HOSTNAME} reachable with inverse conditions and use that for trigger dependencies

Both of these ... well ... lack a certain kind of elegance :-)

Is there any better way?

Thanks!

[Dave96]

I'm also interested in a solution for this problem.

[lukaswoj]

Yeah, me too

Please take a look at this image:


Each item (host) is monitored by icmpping defined as Item in Template. There is also Template Trigger defined for this Item.

I dont know hot to turn ON Triggers beetwen:
2klatka - 5klatka
5klatka - 11klatka
11klatka - 89-98
89-98 - mieszkanie_60

when trigger beetwen mieszkanie_60 - 3klatka goes ON (like on this image).

And now.... I will quote schneck:

Possible solutions would be:

* don't use templates and define every trigger/action for every host
* define an additional trigger PING {HOSTNAME} reachable with inverse conditions and use that for trigger dependencies

Both of these ... well ... lack a certain kind of elegance :-)

Is there any better way?

Thanks in advance for any response
... and sorry for my poor english

[schneck]

After some searching, the only way that I've found is

{mailsrv:icmppingsec.last(0)}>0.0 & {mailsrv:smtp_perf.max(615)}=0

which triggers only if the mail server can be pinged, but hasn't answered SMTP connections for the last 10 minutes.

This is some kind of solution, but still not as good als real dependecies would be (still waiting for that NOT checkbox ... yes, I know it's open source, so if I really need it I could do it myself :-)

\B.