Zabbix Trapper and icmp pinger processes more than 75% busy

[a.naderi]

Hi all, good day.

I have a Zabbix Monitoring server (5.0.3) with server configuration about :

CPU : 2 cores of virtual CPU
RAM : 16GB
HDD : 50GB

I have about 30-40 host for monitor.


And now my server always have problem :

Zabbix trapper processes more than 75% busy:



Zabbix icmp pinger processes more than 75% busy:




Anyone have a solution for this problem?
And what configuration should I used for best database optimize.
I also checked this blog but nothing happened for me:
http://redes-privadas-virtuales.blogspot.com.br/2012/10/zabbix-poller-processes-more-than-75_28.html



Thank you and have a nice day.

[a.naderi]

Update :
I changed StartPingers from 1 to 5 and it's look like fixed completely !
but for Trapper I first changed it to 7 from 5 then 7 and finally change it to 13 ! it's better now but look :




This is my system information :

Zabbix server is running Yes localhost:10051
Number of hosts (enabled/disabled) 55 49 / 6
Number of templates 140
Number of items (enabled/disabled/not supported) 14819 13581 / 983 / 255
Number of triggers (enabled/disabled [problem/ok]) 8049 6572 / 1477 [117 / 6455]
Number of users (online) 3 1
Required server performance, new values per second 102.86

Do you have any suggestions ?

[ardabeyaz]

I have the same problem with zabbix 6.2 and 6.4 (mysql) that started today. I have no idea why it started all of a sudden, and server resources are half idle at peak. Changing trapper and pingers did not affect at all. Anyone else having the same problem ?

NVPS: 40

[Luisbiz]

Same problem here.

[hreyes]

Similar problem here, has anyone been able to resolve this issue?

Thanks

[tim.mooney]

Similar problem here, has anyone been able to resolve this issue?

The person that started this thread did a good job of describing their environment and what their current settings are. You need to provide details like that person did to make it easy for someone to attempt to help you.

The usual answer when someone has a Zabbix subsystem (e.g. icmp pingers, or pollers, or whatever) that is becoming too busy is to SLIGHTLY increase the number of processes related to that subsystem. This does not mean changing "icmp pingers" from 3 to 50 or changing "history syncers" from 5 to 100! It means making a small increase (perhaps from 3 to 5, or maybe if the starting value is 5 change it to 8) and then waiting possibly a few days, to see if the utilization improved and whether it improved enough or whether additional slight adjustments might help even more.

There are situations where the problem is elsewhere (for example, poorly tuned database) and slightly increasing the subsystem that is showing a high utilization isn't going to help (and might hurt), so making small changes and then watching for improvement or a worsening situation helps to make sure you can revert to your original configuration and attempt a different solution.