<%NUMBERING1%>.<%NUMBERING2%>.<%NUMBERING3%> PRTG Manual: System Administration—Monitoring

Available Intervals

Define the intervals available in the drop down list of every object's settings. In the text field, enter one value in each line. Use s, m, h, and d for defining seconds, minutes, hours, and days. By default, there are the following scanning intervals defined:

30s: 30 seconds

1m: 1 minute

5m: 5 minutes

10m: 10 minutes

15m: 15 minutes

30m: 30 minutes

1h: 1 hour

4h: 4 hours

6h: 6 hours

12h: 12 hours

1d: 1 day

Note: We recommend that you do not use intervals shorter than 10 seconds to prevent system overload. Intervals below 10 seconds are not officially supported! The maximum supported scanning interval is 10 days.

You can also define specific points in time to indicate when PRTG executes scanning actions. Enter up to 50 concrete UTC points in time according to the formula

@ UTC hh:mm, hh:mm

Note: Your local time may be different from the UTC time! For more information on how to set specific points in time as a scanning time for sensors, see the More section below.

The unusual detection can set sensors to an Unusual status when there are values that are untypical for the time span in which they are measured. PRTG compares the current average values to the historic monitoring results for this purpose. If the current values show a big difference to the values that are normally retrieved by a sensor, this sensor will indicate this with the unusual status. You can define the granularity of the detection here (this means, how big the difference must be to cause an unusual status). If you disable the unusual detection (both settings to Never), sensors will never show an unusual status.

Note: You can enable and disable unusual detection for specific devices, entire groups, and probes in the respective Object Settings.

Show Unusual When

Define when a sensor shows the unusual status, comparing the weekday. If you enable the detection here, the average of the values which were measured on the day before is compared to the average of the same weekday in previous weeks. Choose between:

• Never: Disable unusual detection for weekday average.
• 24h average is <80% or >120% of weekday average: The average of the values measured on the day before is either lower than 80% or higher than 120% than usually on the same weekday.
• 24h average is <50% or >200% of weekday average: The average of the values measured on the day before is either lower than 50% or higher than 200% than usually on the same weekday.
• 24h average is <20% or >500% of weekday average (recommended): The average of the values measured on the day before is either lower than 20% or higher than 500% than usually on the same weekday.
• 24h average is <10% or >1,000% of weekday average: The average of the values measured on the day before is either lower than 10% or higher than 1,000% than usually on the same weekday.
• 24h average is <1% or >10,000% of weekday average: The average of the values measured on the day before is either lower than 1% or higher than 10,000% than usually on the same weekday.

For example, consider a traffic sensor that usually measures 100 MB average traffic on a weekday. If you choose the first option, it would show an unusual status if the average from the day before is below 80 MB or above 120 MB.

Show Unusual When

Define when a sensor shows the unusual status, comparing the hour-of-day. If you enable the detection here, the average of the values which were measured in the hour before is compared to the average of the same hour on the same weekday in previous weeks. Choose between:

• Never: Disable unusual detection for hour-of-day average.
• Hourly average is <80% or >120% of hour-of-day average: The average of the values measured in the hour before is either lower than 80% or higher than 120% than usually in this hour of this weekday.
• Hourly average is <50% or >200% of hour-of-day average: The average of the values measured in the hour before is either lower than 50% or higher than 200% than usually in this hour of this weekday.
• Hourly average is <20% or >500% of hour-of-day average (recommended): The average of the values measured in the hour before is either lower than 20% or higher than 500% than usually in this hour of this weekday.
• Hourly average is <10% or >1,000% of hour-of-day average: The average of the values measured in the hour before is either lower than 10% or higher than 1,000% than usually in this hour of this weekday.
• Hourly average is <1% or >10,000% of hour-of-day average: The average of the values measured in the hour before is either lower than 1% or higher than 10,000% than usually in this hour of this weekday.

Consider a traffic sensor that usually measures 10 MB average traffic within an hour. If you choose the first option, it would show an unusual status if the average from the hour before is below 8 MB or above 12 MB.

Logging

Define if unusual events will be written to the log file. Choose between:

• Do not log unusual events
• Write unusual events into the log

Similar sensors detection enables PRTG to analyze sensor data for similarities. The detection will run in the background with low priority. The recommended setting for similar sensors detection is to let PRTG automatically decide how many channels will be analyzed. However, you can also override this setting.

Note: When similar sensors analysis is turned off or you have exceeded 1,000 sensors and have chosen the automatic analysis depth option, the similar sensors entry will not be shown in the main menu bar.

Analysis Depth

Define the number of channels PRTG will analyze to detect similarities between sensors or turn the analysis off. Choose between:

• Manage automatically based on sensor count (recommended): The analysis depth depends on the total number of sensors you have configured. PRTG will analyze all channels for up to 500 sensors, and only the primary sensor channels for up to 1,000 sensors. If exceeding 1,000 sensors, the analysis will be turned off.
• Analyze primary channels only: Only the primary channels of sensors are analyzed. Be aware of potentially high CPU load of PRTG when choosing this setting for more than 1,000 sensors.
• Analyze all channels (higher CPU load): Similarity detection is applied to all channels. Be aware of potentially high CPU load of PRTG when choosing this setting for more than 500 sensors.
• Turn analysis off: No similarity detection takes place. Choose this option if you are not interested in the analysis results or you want to keep PRTG's CPU load at a minimum.

With the sensor recommendation engine, you can automatically analyze the devices which you add to PRTG for useful but missing sensors. PRTG can show you sensor types that give you a much more detailed picture about your network when you monitor the respective devices with them. The analysis runs with low priority in background when you add a new device, the last analysis has been 30 days ago, or you click Recommend Now in the context menu of a device or on a device overview tab.

When the analysis is finished for a device, suggested sensor types appear in the Recommended Sensors table list on the device overview tab. You can add them directly to the device with the Add Sensors button.

Note: To recommend SNMP sensors for a device, the detection engine uses the SNMP version that you define in the Credentials for SNMP Devices section of the device settings.

Note: The detection engine cannot run at the same time as Auto-Discovery. PRTG gives auto-discovery a higher priority, for example, when you add a new device, and queues the sensor recommendation in this case.

Note: The detection engine only checks if a certain sensor type currently exists on a device and recommends to add this sensor if not found. It does not check if there already existed a sensor of this type on the device and was deleted previously! Because of this, sensor types appear in the "Recommended Sensors" list again if they are categorized as useful for the according device when you have deleted all sensors of this type.

Detection Engine

Define if you want PRTG to analyze your devices to recommend useful sensor types. Choose between:

• Manage automatically based on sensor count (recommended): PRTG runs the detection engine for installations with up to 5,000 sensors by default. If you exceed this threshold, PRTG disables the detection engine for performance reasons. We recommend you to set this option to not miss any important data about your network without performance impact.
  Note: You can manually start the analysis for a device with Recommend Now in the context menu or on the overview tab.
• Always show recommendations: PRTG always analyzes your devices even if your installation exceeds 5,000 sensors.
  Note: You can manually start the analysis for a device with Recommend Now in the context menu or on the overview tab.
• Turn recommendations off: PRTG does not recommend sensors. Select this option if you have issues with the performance of PRTG or do not want to see this information on device overview tabs. Also you will not have the option Recommend Now in device context menus or on overview tabs.

Run Discovery at

Define the time when PRTG automatically runs an Auto-Discovery in your network if you configured a daily or weekly Discovery Schedule in the auto-discovery group settings. Choose a full hour between 0:00 and 23:00. We recommend you to choose a time when there is little user activity in your network, because auto-discoveries can produce a certain amount of load.

Desired Minimum Uptime

Define which uptime in percent PRTG regards as 100 percent. This setting affects the colors shown next to the sensor icons in reports. Select one of the predefined values between 90 % and 99.999 %.