Your opinion about trigger

[rubendob]

Hi

based on other trigger we use for know if host or device is available or not I have setup another one to latency, do you think is helpful?

{Template_HN_Basico:icmppingsec[,,,,15000].max(#5)}>100 & {TRIGGER.VALUE}=0| {TRIGGER.VALUE}=1&{Template_HN_Basico:icmppingsec[,,,,15000].count(#10,0,"eq")}<100

Thanks!

[rubendob]

Again

Based on explanation of Zaniwoop I have rewrite the trigger

{Template_HN_Basico:icmppingsec[,,,,15000].max(#5)}>100 &{TRIGGER.VALUE}=0| {TRIGGER.VALUE}=1&{Template_HN_Basico:icmppingsec[,,,,15000].count(#10,100,"less")}<100)

What do you think now?

Is ok?

[rubendob]

Final

Here is my final trigger

{Template_HN_Basico:icmppingsec[,,,,15000].max(#5)}>100 & {TRIGGER.VALUE}=0|({TRIGGER.VALUE}=1&{Template_HN_Basico:icmppingsec[,,,,15000].count(#10,100,"lt")}<100)

Any opinion? :S

[rubendob]

once again

hi all

I change again the trigger

{Template_HN_Basico:icmppingsec[,,,,15000].max(#5)}>100 & {TRIGGER.VALUE}=0|({TRIGGER.VALUE}=1&{Template_HN_Basico:icmppingsec[,,,,15000].count(#10,100,"gt")}<100)

a) if TValue = 0 & last 5 periods of time ms was > 100(ms) => TRIGGER CHANGE
b) if TValue = 1 (PROBLEM) & during last 10 peridos, historical value was greater than 100(ms) AND now value is < 100 ms => TRIGGER CHANGE

I think is ok...but not sure..

What do you think?

Thanks

[rubendob]

Hi

My trigger does not work. Do you people has trigger like mine? Mine trigger only fire when some device reach X ms of ping but not returns to NORMAL state when latency goes down.

Any ideas?

[jerrylenk]

Just to be sure I understand what you want:

Condition a) the trigger should switch to 1=Problem if among last 5 ping-times one was >100 (... and it was 0=OK before)
and
Condition b) the trigger should remain on 1=Problem (if it is on 1 and...) as long as any of last 10 pings was >100
-- that is, if 10 pings in a row are good, trigger switches to 0=OK

right?

For that, you need to check condition a) OR condition b)
As humans, we say "and", but to computer logic it's "or"

But you can have the same thing with only on condition, 'cause if one in five samples is >100, it is also in 10 samples, right? That would be:

{Template_HN_Basico:icmppingsec[,,,,15000].max(#10)}>100


If I misunderstood your intention, correct me.

Jerry

[rubendob]

hysteresis

Hi Jerry and thanks comment

Yes, as you said before, will be enough with only one condition. But there is one point I think is little bit annoying in Zabbix, the flap detection and grace period. In fact, I made this trigger based on hystersesis. You know some trigger that check current status of same trigger.

Reads:

http://meinit.nl/zabbix-triggers-flap-detection-and-grace-period

This is useful to avoid continuous warnings of fired triggers. So my goal is to have some trigger based on histersys to detect high latency(iccmpingsec) in net devices but avoid problem of have a trigger continuously activated.

Last trigger I wrote on thread works good at first condition but then nevers recover to OK state. It remains always in PROBLEM state.

So My trigger does not work properly.

Any ideas about how define trigger based on histeresys for high latency?

Example of trigger to avoid flap detection to ping (ping or not ping)

({TRIGGER.VALUE}=0&{test:icmpping[,2,25,68,200].max(#2)}=0)
|
({TRIGGER.VALUE}=1&{test:icmpping[,2,25,68,200].count(#10,0,"eq")}>0)

Thanks

PD: excuse my rusty english :-/