Monitoring large installations?

[Alexei]

Sure. Also the system will be used as a source of screenshots for documentation.

[tdomhan]

Sure. Also the system will be used as a source of screenshots for documentation.

is there any documentation on how to setup large scale enviropnments for 1.6.x?

Is anyone of you guys using proxies for doing the polling work?
If so, does it load a great amount of performance from the main zabbix server?

[Sawn]

Hello

Our setup is managing ~2800 hosts (with 16k items and 8k triggers) the proxies are doing all the work.

We are using 4 proxies, 1 server and two MySQL in replication.

If the zabbix_server does not have to manage a MySQL server load is ok but the proxies and the master database (the two servers in replication) are under heavy load.

[pedmond]

Hello,

I'm planning to activate distributed monitoring on a large installation.

Here is what I already have :

- 1 Zabbix server + 4 proxies + 1 DB that manages 577 hosts, 12230 items and collect 73.79 values per second
- 1 Zabbix server + 4 proxies + 1 DB that manages 1016 hosts, 60682 items and collect 198 values per second
- 1 Zabbix server (which will be the central node) + 1 proxy + 1 DB

My central node will only collect the events.

For the moment, the second DB server really is in heavy load.

[walterheck]

Very interesting thread!

By the way, we are building large ZABBIX setup for pre-1.6 testing purposes (>10K of hosts, 20 Proxies, >100K of items). I am looking forward to nailing down performance-related problems prior to release of 1.6.

Did that environment ever get built? And do you have any documentation on it? These kinds of things are very useful, especially with all the changes that zabbix has seen and the new performance improvements in 1.8.3 and all.