Zabbix - an enterprise-class distributed monitoring system
Posted by Doomshammer
on Saturday, September 22. 2007
at 11:47
in Anwendungen, Arbeit, Computer, English only, Linux/Unix, Privat, Thoughts, Web
I am currently monitoring all my servers with Nagios and I use Cacti to make some fancy graphs of these hosts. As I soon have to disable the server, from which I do the monitoring, I was looking for a new solution. I already read some good articles about Zabbix, so I installed Zabbix on a new server.The installation was fairly simple. Apache + PHP5 + PostgreSQL - quickly compiled and setup. Now download the Zabbix sources from the Zabbix website and compile the server. Copy the frontend into your webroot and perform the rest of the installtion steps - that's it.
The concept of Zabbix is a bit different from Nagios. It it mainly based on it's own agent daemon (which has to be installed on every host that you wanna monitor), where nagios is rather based on some checking tools that are run against the host over snmp or other service ports (check_http for tcp/80/443, check_ftp for tcp/21, and so on).
At the first glance Zabbix is a bit confusing, but after some reading in the manual you quickly get the point about triggers, actions and stuff. What I like about Zabbix is, that it combines Nagios + Cacti + Content Validation. So I can easily montitor all my hosts (Windows, Linux, Solaris ...), generate some graphs and perform website validation. The frontend concept of Zabbix is very mature. I'll show you some brief screenshots to get some overview of it.
The overview page. It lists all configured hosts (currently only 3 as I didn't had the time last night to add all of them). It shows all configured triggers together with a little status box for each host. If the status box is green, everything is fine- is it red... go ahead and check your host
The web content validation screen. You can configure content validation suites for every host. You can even go through a i. e. complete order process on a website and if one of the steps fails, you'll get informed by Zabbix.
The latest data screen. Thanks to the agent that is running on every host, you can easily access lots of host informations w/o having to configure ugly SNMP suites. The client provides lots of data about the host like performance, OS, cpu, network, and so on.
On-the-fly graphs. For nearly every data you can create graphs - on-the-fly. You can define the time period, you can define the look of the graph. You can also create graph-suites which a building graphs automatically. With this tools you can easily graph the cpu load of your hosts, the network traffic and so one.
There are lots of more features with Zabbix. If you want an easy to use but powerful distributed monitoring system, you should give Zabbix a try.
