Zabbix Server 3.4 - Late notification and data graph

[Clontarf[X]]

Interesting. Was your previous OS Redhat/CentOS?

[darklord20k]

Interesting. Was your previous OS Redhat/CentOS?

Previously, I used Ubuntu 16.04 and mysql server, I was used new server to monitor the old, many trigger CPU overload, even I stopped zabbix server service (less often but still sometimes)
I think maybe Ubuntu or mySQL server version's problem (new server use MariaDB) Just my opinion but not sure
I will update later if problems occur again

[darklord20k]

Hi,
I updated to ver 3.4.8, on Debian 9.1, esxi 6, HDD 10000 rpm.
It's work nice for about 1 week, then history syncer busy problem returns I think It's make trigger's recovery dont fire well,
any advise?

[darklord20k]

Closer look

[kloczek]

As usually gaps in graphs data are first symptoms of reaching scale of the stack when passive monitoring (which is default) needs to be switched to active one.
StartPollers=450 With passive monitoring this needs to be increased with NVPS.
Second thing. If this StartPollers=450 it is from server configuration you need to start adding proxies.