-
Problem report
-
Resolution: Won't fix
-
Trivial
-
None
-
4.0.2
-
CentOS Linux release 7.6.1810, all packages updated, kernel 3.10.0-957.1.3.el7.x86_64, postgresql96-server-9.6.11-1PGDG
zabbix-agent-4.0.2-1.el7.x86_64
zabbix-get-4.0.2-1.el7.x86_64
zabbix-java-gateway-4.0.2-1.el7.x86_64
zabbix-release-4.0-1.el7.noarch
zabbix-sender-4.0.2-1.el7.x86_64
zabbix-server-pgsql-4.0.2-1.el7.x86_64
zabbix-web-4.0.2-1.el7.noarch
zabbix-web-pgsql-4.0.2-1.el7.noarch
CentOS Linux release 7.6.1810, all packages updated, kernel 3.10.0-957.1.3.el7.x86_64, postgresql96-server-9.6.11-1PGDG zabbix-agent-4.0.2-1.el7.x86_64 zabbix-get-4.0.2-1.el7.x86_64 zabbix-java-gateway-4.0.2-1.el7.x86_64 zabbix-release-4.0-1.el7.noarch zabbix-sender-4.0.2-1.el7.x86_64 zabbix-server-pgsql-4.0.2-1.el7.x86_64 zabbix-web-4.0.2-1.el7.noarch zabbix-web-pgsql-4.0.2-1.el7.noarch
Zabbix is providing invalid data for Web monitoring items, speeds go down, response times grow (linear) - see attached screenshot.
I have noticed this issue on more than 3 different server installations, both virtualized guests and host variants, issue started right after upgrades of OS packages to latest one (Centos 7.5 -> 7.6) and Zabbix 4.0.1 -> 4.0.2, issues persists with old kernel, without rebooting OS after upgrade.
Not sure if this is related to underlying OS, Intel CPU microcode update or Zabbix 4.0.2 upgrade (there might be a regression in web monitoring code changed in 4.0.2).
See screenshot...
