How do I customize detection for "Unusual" sensor status in PRTG?

We have customers with very different needs. There are many options for customization in order for PRTG to suit your needs.

By default, an unusual detection is enabled. This changes a sensor's status to "unusual" in case there are values measured that are untypical, compared to the historic measurements.

To change the behavior of the unusual detection, or to switch it off completely, please choose Setup | System Administration | Monitoring from the main menu in PRTG's web interface. In the Unusual Detection section, please change settings to your liking.

Note: In PRTG versions previous to 14.x.11 (i.e., version numbers where the last number is lower than 11) you can find this setting under Setup | System Administration | System & Website.

Settings for "Unusual detection" are not that flexible.

This is how "Unusual detection" works:

For example, you got a warning that traffic is unusually high on Wednesday at 9:30 AM. It means that PRTG calculates "1 hour interval average", based on the data obtained on last four Wednesdays, between 9:00 and 10:00 AM, and if traffic is higher than that average, sensor will show "unusual" warning.

It is possible to change the settings for "Unusual detection" (only with predefined values, on the system level). To achieve that, please click on "Setup -> System Administration -> Monitoring" and change the settings in the "Unusual Detection" section. There you can also turn off this kind of alerting (please choose "Never" under "Show Unusual When").

Beside that, you can enable/disable it on device/group/probe/root level under "Settings -> Advanced Network Analysis".

If you want to detect e.g. higher traffic volume than normally or compare it with some value, you can either set channel limits, or you can add volume trigger in the sensor notifications.

For more information about the "Unusual detection", please see this page: Unusual Detection