A basic network monitoring system is a critical part of an organization’s IT infrastructure. The larger the network, the more critical it is to monitor the network for issues. Here is what to monitor and how to decide which resources to monitor:
When it comes to server monitoring, the first property to monitor is uptime. If a server goes down or stops responding, an alarm should be sent out immediately. Most network monitors have a setting to indicate how long or how many times an alarm should run before a notification is sent out. When it comes to server uptime, it should be immediate.
The same level of urgency applies to network switches. These switches connect many users to the network. When anything happens to one, it is quite likely that it will affect as many as 40+ users at a time. Even individual port failure monitoring is useful, given that it could signify the start of bigger problems.
The level of monitoring depends on the importance of the application. Many admins monitor all applications that are part of or have more than on dependency. This is important for two reasons. Firstly, an application failure could result in the failure of several other applications that depend on it. Secondly, it could indicate a problem with an upstream service.
Beyond these three primary categories, other devices to monitor could include VPN’s, Wireless controllers, etc. Only monitor and send out alerts for systems that the organization considers business critical.