As network administrators, we all want our network to be as reliable as possible. If the network develops a problem or is reaching capacity in some area, we want to know about it as soon as possible preferably before users begin complaining. Your network devices will report symptoms of problems by generating events. An event in this context is a message indicating that a device or application in your network has discovered something of note. Your network devices will generate many types of events automatically. In addition, you can use triggers to define or modify the conditions under which events are generated. Triggers are points of interest in specific data objects that generate events when these points are satisfied. A threshold is a type of trigger set on continuous data streams. Now that you've baselined your network and started collecting performance data on that network, you have enough information to make sure that your network alerts you when things become abnormal. The steps you'll need to do this include the following:
The events generated need to be analyzed to determine whether they represent a fault condition or diagnosis of a problem in your network. This analysis is done by an event management system. This system then takes action on faults, such as notifying network operators about the fault, paging you, or just logging the fault to a file for later analysis. The process of analyzing events and processing faults is the subject of the next chapter, "Event and Fault Management." This chapter covers the following topics:
|