Understanding Alerts

The Alert Definition

When you define an alert, you specify the circumstance that causes that alert to occur. This circumstance is the trigger. To define the trigger, you must make the following selections:
  • the range to evaluate
  • the interval in the range in which the data point values must occur
  • the percentage of data points that must occur in that interval

How Alerts Are Triggered

An alert is triggered in the following way:
  1. SAS BI Dashboard checks for a change in the indicator data, updates to the indicator alert definition, or both. By default, this check happens every 3 minutes. A SAS BI Dashboard administrator can change this setting.
  2. If the percentage of data points in the interval has changed, the alert criteria are evaluated.
  3. If the criteria are met, an alert is triggered and a notification is sent to the subscriber.
Note: If the alert criteria are met at the time that you create the alert, the alert is triggered immediately. Afterwards, the alert is triggered only when the data changes and the alert criteria are met.
When alert criteria are evaluated, it is evaluated by subscriber. If a subscriber has multiple alerts, each alert is evaluated for that subscriber before proceeding to the next subscriber. Therefore, if there are a number of subscribers for the same alert, it is possible for one subscriber to receive the alert notification later than another subscriber. The lapse in time might even cause the alert to be evaluated differently. However, a group with multiple members is treated as one subscriber. Therefore, it is more efficient to setup an alert for a group subscriber, than to add multiple subscribers individually to an alert.

Alert Criteria and Data Aggregation

If an indicator can aggregate data, it is important to understand that the percentage that is used by alerting is calculated by evaluating the underlying indicator data to determine the number of data points in the interval. The software does not evaluate the aggregated results presented in the graph view. For example, if a bar chart has three bars, each bar can represent an average of multiple data points. A given bar might fall within the alert interval. However, it is possible that none of the individual data points associated with that bar fall in the same interval. In this case, the three bars reflect the aggregated data, not the data points used in the alert calculation.
The following indicators support alerts and can aggregate data:
  • bar chart with bullet
  • bar chart with reference lines
  • interactive summary and bar chart
  • interactive summary and targeted bar chart
  • targeted bar chart
  • tile chart
  • waterfall chart

Changes That Can Affect Alert Calculations

In addition to underlying data changes, updates to any of the following indicator components can affect the interval percentage calculation and cause the alert to be reevaluated:
  • If the indicator source data is changed, the data available for evaluation can change. For example, if you update the filters defined in an information map, the new filter definition can cause a change in the number of data points that are evaluated.
  • If the indicator data definition is changed, the data available for evaluation can change. The following changes are examples of what can cause the data that is evaluated to change:
    • updating the indicator data definition to modify an SQL query
    • changing the selected items for an information map
    • updating the Group by criteria for a table
  • If the range is changed, the calculated percentage can change. For example, if you change the upper limit of the range from 100 to 200, the percentage of data points that exist in the range and in its intervals can change.
  • If the indicator definition is changed, the data values that are evaluated can change. For example, if you change the Y-axis value setting from Height to Weight, the data that is evaluated is changed.