Zabbix for large environment setups

[jix]

I use zabbix in a large environment.

Number of hosts (monitored/not monitored/templates) 7718
Number of items (monitored/disabled/not supported) 365383
Number of triggers (enabled/disabled)[problem/unknown/ok] 288473
Required server performance, new values per second 861.46

I use 1 zabbix server and 3 proxies. All blades in the same chassis.
All running mysql 5.5 for the backend.

I modified the agent startup script with my own logic for discovery.

This generally works very well, with the exception of the Availability Report.
That chokes to death.

Would you please share your backup solution and if you have any load balance in your database?

[ksb]

Hi!
Which will be the best solution for splitting PostgreSQL database on two raid10 arrays in performance point of view?

[innocent]

Can somebody share an example/module/script on how to automatically add a new kickstarted proxy (for example with puppet) to the zabbix server with the zabbix api?

[Colttt]

@innocent: you can use autodiscovery!

@ksb: what do you mean?

[heaje]

Correct me if I'm wrong, but auto discovery won't add a new proxy into the administration-> DM section. I actually need to do this myself and will be writing a script to do it in the somewhat near future. I'd be happy to share it hear when I'm done (it will be in PHP using the API).

[innocent]

@Colttt, thanks, I will have a look at that.

I do use auto discovery for detecting new agents in our environment.
Will try to make a rule for proxies as well.

[innocent]

@heaje,

Thats what I also thought. If I'm right In an earlier post in this topic somebody altered the proxy init scripts. And used the api to add the proxies to the zabbix server.

As we use puppet to configure/manage our servers in our environment, I want to try to do this with puppet in combination with the zabbix api.

[ksb]

@Colttt
I mean, we have 8 x SSD for Zabbix DB server and thinking the best way to utilise these disks. For now we decided:
1) softraid
2) system on mirror + 2 spares
3) pg data on raid1+0 (16K chunks, as said by Intel and shows best perf. with pgbench)
4) pg xlog on raid1 + 2 spares
It seems okay?

[natalia]

I use a lot of external scripts that capture data so I utilize Zabbix sender.(if you can collect a number of values at once, Zabbix sender is very efficient at doing bulk updates)
I have modified the frontend to show more than the last 20 issues on the dashboard.
I have modified the frontend to reduced the time a trigger stays on the trigger status page when it has returned to an Okay state.

Hi,

Which zabbix version do you have ?
Could you share how your changes for the above ?

Thanks a lot !

[lukemacneil]

Would you please share your backup solution and if you have any load balance in your database?

I don't back up the databases (server or proxies), just the templates. If it croaks, then it will have to be rebuild it. I made sure that was well known to management, though I'm sure they'll forget by the day it fails.

[lukemacneil]

@ lukemacneil : Thanks for the info! Sounds like a great way to do proxy management.
You mentioned something about the API creating the host if it doesn't exist yet.
I assume that is handled by the PHP script you've mentioned?

Correct. The the init script curls a php script, which is what creates the host and assigns the templates based on the environment (test/prod).

Here's the function I use to do that:

function AddHost($host,$ip,$env){
        /*
                This function will add a host to the zabbix database
                based on hostname and IP address. It uses the default 
                group "Production Servers" and "Linux" as the template.
        */

        try{
                $api = new ZabbixApi('http://my_webserver/api_jsonrpc.php', 'myuser','mypass');
        }catch(Exception $e){
                echo $e->getMessage();
                exit;
        }//end try

        $port='10050';
        $useip='1';
        if($env =='p'){
                $groups=array(array("groupid"=>7));
        }else{
                $groups=array(array("groupid"=>6));
        }
        $templates=array(
                "templateid"=>10021
        );

        $result=$api->hostCreate(
                array(
                        "host"=>$host,
                        "interfaces"=>array(
                                'type'=>1,
                                'main'=>1,
                                'useip'=>1,
                                'ip'=>"$ip",
                                'dns'=>"$host",
                                'port'=>'10050'
                        ),
                        "groups"=>$groups,
                        "templates"=>$templates
                )
        );

}//end function 

[emz]

Is using Zabbix Proxy beneficial for the performance of the main Zabbix database?

[kloczek]

Is using Zabbix Proxy beneficial for the performance of the main Zabbix database?

No
Using proxies is about scaling up process of collecting monitoring data from monitored hosts.
Database queries speed/latency impacts process of processing those data on server (evaluating those data against trigger.definitions, processing alarms data, storing monitoring data in history tables, create trends data and few more which are done only on the zabbix server).

[ik_zelf]

No
Using proxies is about scaling up process of collecting monitoring data from monitored hosts.
Database queries speed/latency impacts process of processing those data on server (evaluating those data against trigger.definitions, processing alarms data, storing monitoring data in history tables, create trends data and few more which are done only on the zabbix server).

With current version (v5) it does help. Now you can do high frequency monitoring of items and -for example- filter values so only changes are passed to the server (optionally with a heartbeat)
This allows for quick responses and offloads quite a lot of work to the proxies.

[kloczek]

With current version (v5) it does help. Now you can do high frequency monitoring of items and -for example- filter values so only changes are passed to the server (optionally with a heartbeat)
This allows for quick responses and offloads quite a lot of work to the proxies.

That type of filters are available since 4.8.x.