See this list in the zabbix documentation.
We don't use all of those, so it depends on the roles your server has assigned to it.
We will be monitoring your machine by using a piece of software called the Zabbix agent. The agent is available for many different platforms, listed below.
If you are using any of these Operating Systems, you can usually find a package in a repository that provides the zabbix agent. The advantage of using your packaging mechanism is that updating, removing and installing becomes much simpler. Please check with your distribution to find the best repositories to use.
| Distribution | Repository | Comment |
|---|---|---|
| Debian Lenny | lenny-backports | http://packages.debian.org/lenny-backports/zabbix-agent |
| RedHat/CentOS | Andrew Farley | http://andrewfarley.com/sysadmin/rpm-repository-online / http://repo.andrewfarley.com |
| Other | Check your distro-provider |
Pre-compiled agents for windows can be found on the zabbix website. If you are using any of these Operating Systems, you can usually find a package in a repository that provides the Zabbix agent. The advantage of using your packaging mechanism is that updating, removing and installing becomes much simpler. Please check with your distribution to find the best repositories to use.
We monitor each of the different parameters we keep track of in a different interval. That means for instance that we monitor ethernet traffic every 5 seconds, cpu usage every 10 seconds, but the total amount of memory in your server is only updated once every 24 hours currently. In short, we set the intervals to something that makes practical sense.
We also have two different variables that determine how long we keep the monitored data in our database. We currently keep the history for the past 60 days. That means that we keep every single collected value for 60 days. After that, the data gets summarised and we keep only the trends. We keep those for 365 days.
The amount of system resources our monitoring agent uses on your system depends on the amount of server monitoring the agent has to do. If you have many roles assigned to your server, it will be a bit more busy then with fewer roles.
Here is an example of the running zabbix processes on one of our web servers:
[09-03-11 13:50:09] [root@xyz.com]
[~] # ps aux | grep zabbix
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
zabbix 6394 0.0 0.0 49168 844 ? SN 09:46 0:00 /usr/sbin/zabbix_agentd
zabbix 6395 0.0 0.0 49168 1468 ? SN 09:46 0:09 /usr/sbin/zabbix_agentd
zabbix 6396 0.0 0.0 51264 1468 ? SN 09:46 0:08 /usr/sbin/zabbix_agentd
root 20472 0.0 0.0 5164 780 pts/1 R+ 13:50 0:00 grep zabbix
Our SMS costs are included in your monthly server monitoring price. That means you don't pay a single cent extra for text messages we send to your mobile in case of alerts arising from potential issues on your servers.
For an overview of networks that are covered, please refer to this page on our SMS provider's site
Open Query, founded by Arjen Lentz, develops and delivers training and expertise for MySQL, MariaDB, Percona Server & related technologies. They have engineers on four continents, generally working remotely.