Large Zabbix deployment question

[ataylo13]

I am looking to deploy zabbix in a fairly large environment (roughly 1800 servers). From the initial templates I have created I am looking at ~40 Items and ~20 Triggers per host. I am looking for a way to determine how large the database will become (only storing the data for about 2 months ish) and what polling intervals I can achieve.

I know that these are very broad questions, but i am looking for some warm fuzzies that I have the right hardware before I ramp up completely.

I am also leaning towards going ahead and deploying 1.5.1 for it's distributed polling and db improvements over 1.4.

The servers are RHEL 4U6
The database is dual proc dual core
the zabbix server is single proc dual core vm

Thoughts/opinions?

[nelsonab]

Here was a post from about a year and a half ago.

http://www.zabbix.com/forum/showthread.php?t=3259&highlight=hardware

[BusteR81]

use mysqlsize command on zabbix_agentd...

I am looking to deploy zabbix in a fairly large environment (roughly 1800 servers). From the initial templates I have created I am looking at ~40 Items and ~20 Triggers per host. I am looking for a way to determine how large the database will become (only storing the data for about 2 months ish) and what polling intervals I can achieve.

i have an old Zabbix server on a CentOS 4 system dated back 10 year ago therefore the disk space is only 6 GB. i am still currently using this old system to monitor about 30 hosts. like you, the zabbix mysql dB eats up my disk space quickly ~2mths.

my solution: add a remote "mysqlsize of zabbix db command" on the zabbix_agentd.conf running on zabbix_server. add a item "zabbix database size". have the zabbix_server monitor the database growth size and after a day or 2, then you will know "roughly" how to tune the polling intervals / history period to be kept / trends period to be kept from the graph.