Hello community,
I'm quite new to Zabbix and wanted to ask some questions about few components to better troubleshoot issues.
Here is my architecture (all on CentOS 7 virtual servers) :
- zabbix1 : zabbix "core" server (v3.2.4)
- zabbix2 : zabbix proxy
- zabbix3 : MySQL DB dedicated to zabbix proxy and zabbix server only
It manage I.T. stuff. I have about 450 hosts and 15k items (total including non-working/disabled ones). The fasters checks are for environmental data such as temperature and humidity, or pings (every 1min).
1) Zabbix config :
It seems I can config pollers, trappers, pingers, etc... on zabbix1 AND zabbix2. I only use zabbix2 as a proxy for all my hosts.
-----> Is it useful to start, let's say, 10 pollers on the server and 10 pollers on the proxy ? (5 pollers are not sufficient.)
I tried to disable all the pollers/trappers on the server, but it needed at least 1 trapper to receive data from the proxy if i understood.
2) Zabbix queue :
When I look to zabbix[queue] history (checked every 1min) it return different values than others displayed in the web UI "Queue overview" section.
For example, the last check show 5 items in the queue (and 0 in the "+10min queue") but in the overview i have 300+ items waiting for more than an hour.
zabbix[queue,1,60] seems to display different values as well.
-----> How can I get the same value displayed in the overview ? Is there another item like zabbix[queue] but more accurate ?
3) Zabbix performance monitoring :
I monitor Zabbix performance with Grafana. It works pretty well, but when i started a dedicated dashboard for Zabbix stats, i saw that "busy" graphs looked nearly the same for zabbix1 and zabbix2 servers.
Pollers, trappers, pingers, etc... are not of the same amount on each server. For example there is 1 poller started on zabbix1 (main server) and 15 on zabbix2 (proxy).
The first server is monitored by a "Zabbix server" dedicated template, and the second by a "Zabbix proxy" dedicated template which have few different items.
-----> In my case, is it useful to monitor the same items on zabbix1 and zabbix2 ? Can the little difference between values be caused by the latency between each check ?
I'm not sure if i'm clear enough on what i'm asking, so any explanation can be added if necessary.
Thanks in advance for your help!
I'm quite new to Zabbix and wanted to ask some questions about few components to better troubleshoot issues.
Here is my architecture (all on CentOS 7 virtual servers) :
- zabbix1 : zabbix "core" server (v3.2.4)
- zabbix2 : zabbix proxy
- zabbix3 : MySQL DB dedicated to zabbix proxy and zabbix server only
It manage I.T. stuff. I have about 450 hosts and 15k items (total including non-working/disabled ones). The fasters checks are for environmental data such as temperature and humidity, or pings (every 1min).
1) Zabbix config :
It seems I can config pollers, trappers, pingers, etc... on zabbix1 AND zabbix2. I only use zabbix2 as a proxy for all my hosts.
-----> Is it useful to start, let's say, 10 pollers on the server and 10 pollers on the proxy ? (5 pollers are not sufficient.)
I tried to disable all the pollers/trappers on the server, but it needed at least 1 trapper to receive data from the proxy if i understood.
2) Zabbix queue :
When I look to zabbix[queue] history (checked every 1min) it return different values than others displayed in the web UI "Queue overview" section.
For example, the last check show 5 items in the queue (and 0 in the "+10min queue") but in the overview i have 300+ items waiting for more than an hour.
zabbix[queue,1,60] seems to display different values as well.
-----> How can I get the same value displayed in the overview ? Is there another item like zabbix[queue] but more accurate ?
3) Zabbix performance monitoring :
I monitor Zabbix performance with Grafana. It works pretty well, but when i started a dedicated dashboard for Zabbix stats, i saw that "busy" graphs looked nearly the same for zabbix1 and zabbix2 servers.
Pollers, trappers, pingers, etc... are not of the same amount on each server. For example there is 1 poller started on zabbix1 (main server) and 15 on zabbix2 (proxy).
The first server is monitored by a "Zabbix server" dedicated template, and the second by a "Zabbix proxy" dedicated template which have few different items.
-----> In my case, is it useful to monitor the same items on zabbix1 and zabbix2 ? Can the little difference between values be caused by the latency between each check ?
I'm not sure if i'm clear enough on what i'm asking, so any explanation can be added if necessary.
Thanks in advance for your help!