mysql problems on proxy

[LenR]

After our migration to Zabbix 3.4, one of our proxies will have problems after a few days. I found it was short on memory and the mysqld process was using a lot of memory.

It looks like these mysqld.log messages start about the time the problems start:

18 [Note] Aborted connection 18 to db: 'zabbix' user: 'zabbix' host: 'localhost' (Got an error reading communication packets)

Zabbix 3.4.3
mysql-community-server-5.7.19
Red Hat Enterprise Linux Server release 7.4 (Maipo)

The proxy is a VMWare VM, 4G ram, about 1000 win/linux hosts, mostly zabbix agent(active) items, maybe 10-20% snmp items, 250 web tests and the typical tag-on simple checks for 950 NVPS. This isn't are busiest proxy but the bigger ones are 95% snmp items.

For now, I just stop the proxy and mysqld, then restart both via cron daily.

Any ideas?

[kaspars.mednis]

Hi !

Are there corresponding error messages in zabbix_agentd log files ?

What are the following MySQL settings:

wait_timeout
max_allowed_packet

SHOW GLOBAL VARIABLES LIKE 'wait_timeout';
SHOW GLOBAL VARIABLES LIKE 'max_allowed_packet';

Regards
Kaspars

[LenR]

wait_timeout = 28800
max_allowed_packet = 4194304

There is nothing that appears to be related in the proxy log, just typical unsupported items, vagrant hosts ans an occasional slow query log during housekeeping.