Set thresholds

Thresholds are markers that indicate certain levels have been reached. Many component monitors used in SAM application monitors have thresholds that can be adjusted and set to act as trigger points. For example, if monitoring CPU usage, a Critical threshold set at 90% would be typical. You can use this threshold as a trigger for SAM to send you an email alert when a Critical status threshold is breached.

SAM includes predefined static thresholds for most statistics, but you can override these thresholds and customize them on a per-object basis. In the Orion Web Console, the color of a statistic indicates if reached either of two threshold levels: critical and warning. A value that crosses a warning threshold appears yellow, and a critical threshold appears red.

To change the predefined value for a threshold, you can add a static threshold or let the system calculate a dynamic baseline threshold.

  • Static threshold. A constant value set for your threshold. For example, the warning threshold for response time might be 500 ms, and the critical value might be 1000 ms. You should be familiar with the performance of that object to know what a reasonable value for a static threshold is.
  • Dynamic baseline threshold. Data is collected for a week, and then used to calculate mean and standard deviation. The warning and critical threshold values are defined as 2 and 3 standard deviations above the mean, respectively. For example, if the mean value for packet loss for a specific node is 0%, the warning threshold for packet loss would be 3% (+2 standard deviations) and the critical threshold would be 4% (+3 standard deviations). Dynamic baseline thresholds are the most accurate way to define thresholds for a specific device.

Baselines are calculated once, after a week of data is collected. You can recalculate baselines on demand.

To learn more about thresholds, see: