When you have a proper monitoring tool installed and set up on your network, it is up to you to configure it to alert you in the ways you believe are most suitable to you. When the network main server, or servers, is down, a service has been switched off, or a router has become unavailable, the network monitoring tool will send an alerting message to its administrator. Alert messages could be done in a variety of ways depending on what methods the tool itself supports.

Some tools provide SMS (Short Message System) text message alerts sent to the administrator's cellular phone. Other tools send an SMTP email message to the administrator's email address. Other alerts may be done using the following: HTTP(S) POST and HTTP(S) GET actions, or Wake on LAN (WOL). There is also a sound and voice call option in the alerting subsystem of monitoring software.

If you're the type of network administrator that dislikes sound and prefers textual alerts, then you may turn off that option altogether in most software applications.

The frequency of alerts and their robustness depends on both the software capabilities and how the admin configures it to be. In principle, there are alerting rules to be set and Dependencies to be assigned.

From its name, "dependencies" describe the case when the failure of a service device or activity directly or indirectly affects the activity of another device or service. For instance, a particular server on the network provides 2 different services: A and B. If the server itself becomes "down", then we may directly conclude that both services A and B have become unavailable since they are dependent on the server. Therefore, the alert triggered may be used to deliver the whole status message.

Taking this example a step further, we can see that once a major server is down, the number of "failed", "down" and "unavailable" responses to the software system becomes overwhelming. If the system responds to each incoming alert trigger by sending an SMTP, SMS, ICQ or MSN message or even calling the "on call" person in charge, then the person would be literally driven nuts. He or she will therefore be incapable of understanding the problem (or problems) and unable to respond with the appropriate actions.

Therefore, it is the usual practice to configure the network monitoring software to send an alert the second time a mal occurrence takes place, rather than the first.

Moreover, some software allow grant you the flexibility of customizing the alert messages of particular devices so that the content of several alert messages may be aggregated and modified into a single alert message.

Software applications perform ordered checks on the list of devices and services that the network administrator configures into it during setup. This list of entities (with a defined Host name and ID) may be grouped and even changed later on as well.

Sometimes there are devices that you would like to make the software aware of but not yet available for checks. Applications such as Servers Alive help you do so by assigning an "Active" or "Maintenance" mode to our entry. Checks are performed on the entities with an "Active" mode flag on it. This way, the alerts coming from unavailable devices or entities that are undergoing maintenance could be avoided altogether.

Check cycles are then performed on the entity list, one by one, beginning at the top. This is how software like Servers Alive checks each entity status and activity level. Note that the default order by which the entities are checked is by host. In another attempt to reduce the number of incoming alerts to the network administrator, you may wish to configure the software to conduct the checks periodically every X seconds or X minutes depending on your business needs.