Server is unreachable [only after 5 min]

[ittec]

Hi

we are experiencing some problem with trigger "Server is unreachable". When a host is very very busy, by some reason Zabbix Server can't stablish connection with Zabbix Agentd during some seconds then trigger is fired and action starts, sending SMS and Email.

I'm finding some way to continue monitoring of host falled but I want that only alerts when some reasonable kind of time has gone.

Maybe adding more conditions to standard trigger?

{Template_Linux:status.last(0)}=2&{Template_Linux: status.min(120)}=2

Some idea guys?

Thanks

[MrKen]

This is what I use:
({ZabbixServer:agent.ping.last(0)}#1)|({ZabbixServer:agent.ping.nodata(180)}=1)

It seems to work ok, however, I wonder whether I should replace the OR operator with an AND operator. But, as I say, it works. You can change 180 to 300 (5 minutes) or whatever.

MrKen

[ittec]

Hi MrKen

({ZabbixServer:agent.ping.last(0)}#1)|({ZabbixServ er:agent.ping.nodata(180)}=1)

thanks by your reply. One question:

1. Why do you use "#" on the first part of trigger? It's a mistake or has a real meaning?

So if your query works, maybe I can use it to another triggers with the same intention (Avoid false warnings)

Template_Linuxroc.num[httpd].last(0)}<1&{Template_Linuxroc.num[httpd].delta(180)}<1

2. it can be changed to...?

Template_Linuxroc.num[httpd].last(0)}<1&{Template_Linux:proc.num[httpd].nodata(180)}<1

Thanks

PD:Im testing your rule and I think is working well.

[MrKen]

1. Why do you use "#" on the first part of trigger? It's a mistake or has a real meaning?

# means Not Equal to

See the latest manual p.123 Section 4.13.1 Expressions for Triggers.

MrKen

[subba5678]

Hi Mrknen,
Can you please help me what is the exact difference between the below 2 templates . I want to monitor Zabbix Agent . Please help me out which one i need to use.

{Template_Windows:status.min(300)}=2

{ZabbixServer:agent.ping.last(0)}



Thanks,
Subbu

[MrKen]

I shall answer in reverse order.

{ZabbixServer:agent.ping.last(0)}
If ping is successful it will return a value of 1, if ping fails the value is 0. So, if the last value is 0 (zero), the trigger will fire.

{Template_Windows:status.min(300)}=2
The use of 'status' for host availability is not recommended, says Alexei, so I don't use it at all! See here: http://www.zabbix.com/forum/showpost...59&postcount=4

For availability of zabbix_agent, I monitor 'agent.ping' and also 'icmpping'. I set the trigger on icmpping like this:
({Host:icmpping.max(300)}=0)|({Host:icmpping.nodat a(300)}=1)

This means that if the maximum value is 0 (ping fail) for 5 minutes OR there is nodata for 5 minutes, alert me.

MrKen

[EnigmA-X]

Hi Mrken,
Thanks for the reply , I want to use the maximum expression to monitor any service (0r) availability , for that i am using the below template , but I am getting the false alerts from them , Please help me out ,

{ATL640erf_counter["\Processor(_Total)\% Processor Time"].max(3600)}>90

Thanks,
Subbu

I would like to kindly suggest that you start reading here and stop hijacking forum topics from other users, as this is really not polite

Please do not expect that other people will troubleshoot all your issues, without trying to solve them yourself. For what it's worth, the syntax of your trigger (not template) looks wrong to me.

[subba5678]

Hi EnigmA-X,

Before posting the forum i tried in different ways to solve the issue by myself , If any one seeking for the help means , he tried in all the ways and finally not able to fix the issue . then only people will depend on the Forum . Forum is the place where we can find solutions to all the problems , if its not why we require Google , search engines e.tc , I am not hijacking any forums , i am trying to see what all the issues user are facing while using zabbix , I alredy gone through the Zabbix Manual. According to me Zabbix is a Very Good Tool in the monitoring , we already put the zabbix in production , my intension is not get any down time . If you are Intrested and ready to help please help me ,I hope and thanks for understanding .

Thanks,
Subbu

[EnigmA-X]

Subbu,

Ok, for a simple start: with hijacking forum topics, I mean that you are putting *your* questions, in the topic from *someone else*.

If you have a problem or a question, please start a new forum topic. This keeps the forum clean, readable, but even more important: searchable for other people with the same questions/problems.

So...please put your question in a new topic and I'm sure there are plenty of nice people willing to help you!

[subba5678]

Hi EnigmA-X,

I am posting my queries in the other because i already tried with all the procedure which was updated in that forum but still i am not able to find it out , so that is the reason i posted there , I agree with your point to start a new topic , if some one as already post their views in one forum , there might be a chance the other people wont respond to that , as the solution was already posted , if there no posts related to that query then i will post a new thread ,I posted here because i found my issue is closely related to this post . If you find solution for the below issue . please help me


{TAL720erf_counter["\Processor(_Total)\% Processor Time"].max(3600)}>90

Thanks,
Subbu

[drumspirit]

For availability of zabbix_agent, I monitor 'agent.ping' and also 'icmpping'. I set the trigger on icmpping like this:
({Host:icmpping.max(300)}=0)|({Host:icmpping.nodat a(300)}=1)

For me, the trigger nodata(300)=1 doesn't work.
If I set it to "1", the trigger doesn't fire at all. If I set it to "0", the tiggrer fires after 30seconds.

[ingus.vilnis]

Hi,

To check availability of zabbix_agentd you can also use Zabbix internal item with key:

zabbix[host,agent,available]

This item will return you 1 when available and 0 when unavailable which also makes the trigger pretty simple:

{<host or template>:zabbix[host,agent,available].last(300)}=0

You can read more on internal items here: https://www.zabbix.com/documentation...types/internal

Best Regards,
Ingus