I am rolling out a new deployment of Zabbix 3.4. In my environment, we are placing a zabbix proxy in each of our data centers and having them do all the data collection thus leaving the main zabbix app server (zabbix-server) to do all the trigger evaluations and actions(email, scripts, etc...) I chose this design so that no one data-center can overload the zabbix server process and slow data collection for every other data-center. With that design in mind, here is my problem (or perhaps just a misunderstanding of how internal metrics are calculated.)
I am still in the testing phase of this, so at the moment, I only have a single standalone zabbix-server (physical server) and a standalone zabbix-proxy (physical server) running collecting data for a single data-center. When I look at the "zabbix[queue]" value on the zabbix-server and the same value on the zabbix-proxy, I would expect to see different values. (Note that the zabbix-server IS collecting some of it'e own metrics where as the Proxy is collecting substantially more) I have even go so far at to disable all metrics not collected by the proxy (except for the zabbix-server queue metrics) and was still seeing identical queue number for both the server and the proxy.
Am I doing something wrong? am I missing the obvious? I know there is a finite number of metrics zabbix can collect per-second before you start seeing data collection/processing delays.
Carl S
I am still in the testing phase of this, so at the moment, I only have a single standalone zabbix-server (physical server) and a standalone zabbix-proxy (physical server) running collecting data for a single data-center. When I look at the "zabbix[queue]" value on the zabbix-server and the same value on the zabbix-proxy, I would expect to see different values. (Note that the zabbix-server IS collecting some of it'e own metrics where as the Proxy is collecting substantially more) I have even go so far at to disable all metrics not collected by the proxy (except for the zabbix-server queue metrics) and was still seeing identical queue number for both the server and the proxy.
Am I doing something wrong? am I missing the obvious? I know there is a finite number of metrics zabbix can collect per-second before you start seeing data collection/processing delays.
Carl S