Stopping false alarms for simple check

[predatorz]

Hi,

I am using the following expression for simple check.
{Template:icmpping.sum(300)}=0
{Template:http_perf.sum(300)}=0
{Template:smtp_perf.sum(300)}=0 and other simple check(ssh_perf, tcp, http).

i have tried many expressions hoping to stop the false alarms but to no avail
Ping interval=60s
{Template:icmpping.max(#5)}#1
{Template:icmpping.max(#5)}<1 | {Template:icmpping.max(#5)}>1
and others

The trigger works well to inform us(Email) when service is down (Trigger becomes true, when data collected equal 0 for 300s). However, when zabbix get back value of 2(timeout from the check), it will set trigger back to false(But service is still down), inform us again(Email). after 5 mins, zabbix get data equal 0, it inform us again. The process repeats over and over. Resulting in a lot of false alerts and emails.

Anyway that i can change the trigger expression to keep false alerts to minimal or stop it? How does zabbix determine time out? Anywhere in zabbix to modify?

Hope to find a solution.

[predatorz]

Anyone can help?
My brain is drained.

[predatorz]

Bump.. Help

[predatorz]

Any hints/help???

[predatorz]

Does anyone knows whether simple check functions like pop_perf, smtp_perf, http_perf, ssh_perf and others return value of 2 when timeout?

[Alexei]

It seems it is not possible to create a trigger which would be true if all values are either 0 or 2 for a period of time currently. A new function, valcount() or count(period,value) will be introduced soon (in 1.3.x, pre 1.4) in order to address this issue.

[predatorz]

Thanks for the information, but does the new function help to reduce false alerts?

[predatorz]

Using {template:icmpping.sum(#5)}=0 and other similar simple check, problem is after it is triggered, even a value of 2(timeout) will set trigger back to false even though service still down. Any more modifications i can add to the expression to reduce false alerts?

[predatorz]

Anyone care to shed some light ? Help...

Does those perf items return value of 2 when timeout like pop,smtp,http and others?

[predatorz]

Will it work if a trigger is written to check for host up and it depends on the trigger that check if the host go down?

I doubt it will work. LOL. Most probably, the first trigger will not be able to reset back to false if i write it like above.

True?

[predatorz]

Will it work if a trigger is written to check for host up and it depends on the trigger that check if the host go down?

I doubt it will work. LOL. Most probably, the first trigger will not be able to reset back to false if i write it like above.

True?

AM i correct in this ?
Seems like for those *_perf items, if it the item timeout, value returned is >5, is this correct?

[Alexei]

FYI New function, count(period<,value>), has been implemented. It is already in the latest code. It works with items of any type.

[predatorz]

Yup, saw that in the latest manual.

[Alexei]

Yup, saw that in the latest manual.

I don't think so. The latest manual is on my laptop Function 'count' existed since 1.0, now we just added second parameter.

[rodneyalamarboys]

hello

I have the same problem in my lan because the time out, I don't know how resolve this problem to false alerts