Zabbix poller processes more than 75% busy

[micho]

I tried to play with zabbix config but no luck. My problem is "poller data collector process"


If not listed any other settings is default:
StartPollersUnreachable=50
StartPingers=150 tried with 300 too
VMwareCacheSize=256M
CacheSize=2G
System information

Yes	localhost:10051
230	81 / 9 / 140
5651	3393 / 2164 / 94
2557	931 / 1626 [2 / 929]
5	1
152.26

Also increased max_connections to 256 in mysql config file

[Markku]

How about increasing StartPollers? Default is only 5 on Zabbix 4.4.4.

Btw why is VMware cache increased even though there is no VMware pollers apparently running?

Markku

[micho]

Very weird, tonight out of the nowhere it started working normally. I have not changed anything. What I have noticed was that since I use alter system using email I was getting bunch of emails with triggers since there was circuit outage where the server is connected to. Yesterday I also optimized all my hosts and limited number of items and triggers and graphs but after that change nothing changed. Looks like that was connected somehow to the queue of alerts somehow.

today:

[micho]

Today