Advisor gives appropriate advice and interpretation of observed system behavior based on the following techniques:

• Identification of top instances - Advisor locates the particular Oracle sessions, or UNIX and Linux users (as well as other categories) that contributed most to the total system loading, and shows their individual and combined pattern of activity over time.
• Mean value versus thresholds - The observed behavior is categorized as good, fair or poor. Alerting triggers can be raised if the interpretation is Fair or Poor. An explanation of the consequences of Fair or Poor behavior is offered.
• Variability around the average value - This shows the extent to which the individual measured values of a particular metric are different from the average. Behavior is categorized as fairly constant, rather variable and very variable. Alerting triggers can be raised if the variability is rather or very variable.
• Proportion of time within threshold ranges - This shows the total time when the value of a metric was in one of three possible states: lower than a warning threshold, between a warning and an alarm threshold, or higher than an alarm threshold. The behavior is categorized as Good, Fair or Poor depending on the proportion of time in each state. An explanation of the consequences of Fair or Poor behavior is offered.