Understanding Server Performance Monitoring
Monitoring Server performance is one of the functions of management that is useful for analyzing whether the network is still quite reasonable for use or need additional capacity.
Principles and the workings of the Monitoring Server performance
A system of monitoring the performance of the server performing the process of collecting data about himself and do an analysis of the data with the aim of maximizing all resources. The collected data in General is a real-time data, whether the data is retrieved from the system of hard real-time systems as well as soft real-time. A real-time system is a system where the time required by a computer in the external environment to provide stimulus is a thing that is vital. Time in the sense means that the real-time system runs a job that has a time limit (deadline). Within the time limit an occupation may be resolved correctly or may be unresolved. Real-time systems require that a job should be resolved properly. Something bad will happen if the computer is not able to produce timely output.
Monitoring Server Performance Tools
Monit didn't just monitor your servers, but also sought to fix the problem by taking actions that have been set for certain situations. For example, if the database server crashes, Monit can automatically restart the service if this is the action you want to do.
When you have a group of cluster machine, it's hard to see how this whole cluster at once. Ganglia, presents an overview of the whole cluster. This is a great tool for those of you who work with server cluster, ignoring that these tools are too "exaggerated" for one machine only.
Munin is present with monitoring system based on graph metrics. Can automatically generate a graph of the performance of daily/weekly/monthly/yearly reports and important. monitors core system resources, such as memory, disk space, CPU usage, application servers such as MySQL, Apache, and Squid is simple yet functional tools, by simply adding a few lines of code, you can write a plugin to to monitor almost anything. a good choice for things related to graph the performance of the server, such as the number of users or the popularity of the site.
Tools this one is more like Munin to various things, what makes it different? Cacti is allows you to do the resizing on the graph and see data more clearly and as you like.
Nagios is "industry standard in IT infrastructure monitoring". at least that's what it says on their website. Nagios is indeed a bit complicated to install and configure, but the wealth of features unmatched by the tool on the market and intended for a network administrator who experienced IT. Nagios support monitoring multiple hosts and can send alert via email, pager (if you still use ancient technology) or text messages. Such as Monit, can also be configured to automatically to respond to the problem.
ZABBIX is a tool that was full of features. has a great visualization support including personal settings view, zoom, and mapping. It can send alert via email, SMS or Instant Messaging. These tools also give warnings, which can be useful when you are physically near the machine monitoring.
Observium directed to Linux, BSD and Cisco network. supports automatic discovery of Your network infrastructure, find the network that might be interested in monitoring, these features can be compared with how Wi-Fi software you automatically find the signal in range that can be affordable. Observium provides detailed charts, and can be arranged in addition to nagios to provide warnings. It also integrates well with Collectd (shown below) for a more powerful interface.
Zenoss is the open source version of server monitoring tool, Zenoss Enterprise commercial entirely written in Python. It supports Nagios plugins format, so many that are on the nagios Plugins can be used in the Zenoss. One of the main highlights of Zenoss is a powerful but simple to use user interface.
Similar to Collectd Munin and Kaktus that focuses on a graph metric system. specially designed for performance and portability, this means suitable to work for large systems, low-end, and embedded systems. Designed to use low resources. Collectd which can collect data every 10 seconds without disrupting Your server process, provides statistics in high resolution. You can write additional extensions using C, Perl or Java.
Argus focuses on monitoring network service, and supports IPv4 and IPv6. Have a nice warning procedure. for example, the system detects a problem then the Argus will send a warning and if the problem is still not resolved within the specified amount of time (perhaps because the admin system at home enjoying sleep), another warning will be sent to other people.