Server Monitoring

Question

System administration

Good afternoon. Interested in whether there is a centralized solution for monitoring servers free, preferably open source. The servers are running on win2003/2008. One of the mandatory requirements is to obtain information on all sensors working via snmp for each server. Also interested in the presence of notifications, alerts and other similar functions. What would you recommend to use?

Answer 1

This is not the place to start:
1. Make a list of your infrastructure. what do you want to monitor right now, what do you want to monitor later. Think beyond servers, but also about network equipment, power supply systems (UPS), telephony, and so on.
If you want to work purely on SNMP, then you can get by with zabbix, but keep in mind that many sensors will need to be added to templates. For example, templates for Hewlett-Packard Proliant in zabbix contain fewer control values ​​than check_hpasm returns (this plugin was written about here ).
There are a lot of plugins for nagios/icinga/shinken (see supported hardware and services here - exchange.nagios.org/directory/Plugins and here - www.monitoringexchange.org/inventory/Check-Plugins).
zenoss stands apart - all its plugins are collected in Zenpacks written in python, but it also supports nagios plugins.
2. How many sites does the equipment cost? Is there (or will be required in the future) distributed monitoring?
You need to think about this especially in order to calmly expand the monitoring system in the future, and not change it to some other one.
If the channels between the sites are fat (at least 100 Mbps), you can put anything.
If the channels are narrow - 1-2Mbit, and even ADSL, then install nagios / icinga / shinken. Zabbix is ​​a friend who loves thick channels, but icinga uses them sparingly, in addition, shinken was originally sharpened to work in distributed systems
After you have decided on all this, then it will already be possible to choose what you need.
There are four ways to collect information from servers in nagios/icinga/shinken
1. Through the NSClient++ agent (NRPE requests)
2. Without an agent - Through a raised SNMP service
3, Without an agent - Through WMI (if the server is running under Windows 2003/2008)
4 Agentless - Information about the state of the hardware itself can be obtained through the server management interfaces (HP iLO, IBM RSA, Dell DRAC, etc.)

Answer 2

I also recommend nagios. I have been monitoring servers since 2003 and 2008. I get complete statistics: disk space, temperature, memory, uptime, cpu, etc.

Answer 3

Zabbix can also monitor Windows servers very well.

Answer 4

1. Total Network Monitor is an excellent free solution with a lot of functionality. Spinning under Windows;
2. If you need something to spin on nix, then Nagios ;

Answer 5

newrelic.com/product/server-monitoring is another option. There is a free tariff, a very beautiful thing, easy to install.

Answer 6

I see that there are Nagios offers - but you can also look at Ininga under the niks.
It is generally similar to Nagios - but there are a number of differences.
You can find many articles about both products on Habré, here is a comparison and from the off site :

Answer 7

I am also for I c inga, I especially like the new version of the web interface.