Hi, first i want to thank the developers behind Zabbix for such an excellent software. Now the other reason why i'm writing this post is to ask for soome help in creating a trigger that checks if some devices are alive using a ping. I have a zabbix server running the 1.6.5 version and it works great. It checks the alive status from my other devices using the fping command and i created a trigger based on it that activates if the last value is 0 ({xxxxx:icmpping.last(0)}=0). What i would like to do is to ask more than once before activating the trigger, let's say i take the last 10 values from the ping check and if the check returned 1 less than 6 times it activates the trigger. I did this with something like this {xxxxx:icmpping.count(300,1)}<6.. In this case using seconds (300s) since i couldn't find a way to tell it to use the last 10 values and my zabbix checks every 30s for the status of the hosts. Is this the right approach or is there another better way to do this?Thanks for the answer i may get.
-
-
Hello.
As ping is a binary value, 0 or 1, you can use the sum.
Your trigger can be :
({xxxxx:icmpping.sum(#10)}<6)
It will make a sum on the last 10 values, and test if the result is less than 6.
Have a look to this topic too, where we work on a similar problem :
-
Hi the other thread help me a lot. I got it workin using the next expression
In plain words, i check if the device doesn't answer for the last 6 times (about 3 min) and activate the trigger if this is true. The trigger deactivates when the device answer for at least 2 times.Code:{SERVER:icmpping.sum(#6)}=0 | ({TRIGGER.VALUE}=1 & {SERVER:icmpping.sum(#2)}#2)Comment
-
I'm happy to have helped you.
Comment
-
Thanks fcge! No i just need one more thing to have my zabbix server running perfect and put it in production at my work. And the thing is that sometimes when the server checks some devices (using the icmp check) the return value is Down(0) when it really isn't (since i check in the console and the device answer with both ping and fping), this is quickly corrected when it checks again and it shows Up(1), which is the real state of the device. My problem is why it shows that down value when it really isn't? I'm trying to solve this because i use the availability reports and don't want it to mess the real values, for small it may be. I doubled the time between checks but i get the same result, sometimes it gives me the down value. I don't check too much devices, about 56. What do you recommend? Sorry if i made any mistake in my writing, english is not my natural language
I don't know if i should open another thread with this subject, in any case let me know.Comment
-
Hi, going through my log file i see that zabbix executes /usr/bin/fping -c3 2>/dev/null </tmp/zabbix_server_7595.pinger. How can i change this so zabbix uses only fping <xx.xx.xx.xx>?? Because the thing is that the devices that are giving me problem sometimes they drop some ping packets but this is normal, the thing is that i don't want zabbix to be so sensible with this, and by just checking that the server is alive it gives me an Up(1) value.Comment
-
English is not my native language too, so we'll try to understand each other, and I think I won't see your mistakes.Sorry if i made any mistake in my writing, english is not my natural language
Unfortunatly, I've the same problem. Sometimes, the device is seen as down by Zabbix whereas it's not. That's why I use a trigger on the 3 last values. If 0 pîng successfull : alert (3 ping fail means a problem I think ^^).And the thing is that sometimes when the server checks some devices (using the icmp check) the return value is Down(0) when it really isn't
I think the reports are about triggers and not items values ? So if you also use a trigger a bit less sesitive, it could be ok for the reports no ?
Perhaps other users of this forum have had the same problem ? Do they find a solution ? We can wait for an answer of other people...
Mmmm I'm afraid it's hardcoded in the soft... I'm not sur of that, but I think we can't changing it whithout changing the code ?/usr/bin/fping -c3 2>/dev/null </tmp/zabbix_server_7595.pinger. How can i change this so zabbix uses only fping <xx.xx.xx.xx>??
Perhaps a specialist can confirm it or give us a way to proceed ?
Perhaps can you do that using a external check, with a script which make a simple fping and giving 0 or 1. You just write a small scrit doing thaht... Or using ping -c 3 and a count of the received packets : if it's more than 0, you return 1 (up)... ?the thing is that the devices that are giving me problem sometimes they drop some ping packets but this is normal, the thing is that i don't want zabbix to be so sensible with this, and by just checking that the server is alive it gives me an Up(1) value.Comment
Comment