Counterpart to nodata()

[Andreas Bollhalder]

Hello

I have a trigger which change to ON when there has been no data from the ZABBIX agent for 180s:

{template:agent.ping.nodata(180)}=1

Now, I'd like to have it extended to get OFF only, if there has been data for about 300s to prevent flapping and to keep up with the dependencies:

({template:agent.ping.nodata(180)}=1)|(({TRIGGER.VALUE}=1)&({template:agent.ping.data(300)}=1))

Unfortunatly, there is no function data() itself and I can't use nodata(300)=0. How can I archive this ?

Thanks in advance
Andreas

[Alexei]

You may consider using function count() with one argument, count(180).

[Andreas Bollhalder]

Hello Alexei

You may consider using function count() with one argument, count(180).

If I understand you right, the trigger should looks like:

({template:agent.ping.nodata(180)}=1)|(({TRIGGER.VALUE}=1)&({template:agent.ping.count(300)}>x))

where "x" should be smaller then "300 / interval" (smaller, because I observed that sometimes a value get missed) ?

Thanks in advance
Andreas

[Alexei]

Hello Andreas,

Yes, this is exactly what I meant.

[Andreas Bollhalder]

Hello Alexei

I implemented the trigger like:

({template:agent.ping.nodata(180)}=1)
|
(({TRIGGER.VALUE}=1)&({template:agent.ping.count(300)}[COLOR=Red]<[/COLOR]9))

but it doesn't behave as I expected. I use "<9", because I have the interval for the item "agent.ping" at 30s (in the post before I wrote it accidently wrong as ">9").
Then 180s after stopping the agent, the trigger goes to TRUE and all dependend items goes to UNKNOWN. But when I start the agent (I waited more then 5 minutes), the trigger and it dependend items returns immediatly to FALSE instead after 300s. As far I can see, there aren't 9 values of "1 (Up)" in the the last 300s after starting the agent.

Any ideas ? Do I miss something ?

Andreas

[Andreas Bollhalder]

I'm still unable to get the hysteresis working for the "agent.ping". I tried:

({template:agent.ping.nodata(180)}=1)
|
(({TRIGGER.VALUE}=1)&({template:agent.ping.count(300)}<9))

({template:agent.ping.nodata(180)}=1)
|
(({TRIGGER.VALUE}=1)&({template:agent.ping.sum(300)}<9))

The trigger for the agent ping goes immediatly to FALSE when the ZABBIX agent is reachable again.

Did anyone got a solution for this ?

Andreas

[pbergdolt]

Hello, did you try something like this:


((({TRIGGER.VALUE}=0)&({template:agent.ping.nodata (180)}=1)))
|
((({TRIGGER.VALUE}=1)&({template:agent.ping.count( 300)}<9)))



((({TRIGGER.VALUE}=0)&({template:agent.ping.nodata (180)}=1)))
|
((({TRIGGER.VALUE}=1)&({template:agent.ping.sum(30 0)}<9)))

[Andreas Bollhalder]

Hello, did you try something like this:


((({TRIGGER.VALUE}=0)&({template:agent.ping.nodata (180)}=1)))
|
((({TRIGGER.VALUE}=1)&({template:agent.ping.count( 300)}<9)))



((({TRIGGER.VALUE}=0)&({template:agent.ping.nodata (180)}=1)))
|
((({TRIGGER.VALUE}=1)&({template:agent.ping.sum(30 0)}<9)))

Thank you for your answer. I'm using the following expression without "{TRIGGER.VALUE}=0" (it should be redundant) and it works:

({template:agent.ping.nodata(150)}=1)|(({TRIGGER.VALUE}=1)&({template:agent.ping.count(120)}<4))

I think it has been a problem of the ZABBIX version I used at that time.

Andreas