PRTG Manual: Alerts
Alerting is an important part of monitoring that informs you when there are issues, when values exceed thresholds, or when a sensor status has changed, for example. PRTG offers many ways to alert you about your monitoring data like internal sensor alerts, limits, and lookups. You can also create custom alerts and define notification triggers as extensions to alerts. If a sensor does not have preconfigured limits, you can also configure them accordingly so that you are notified when something is wrong.
Some sensors display internal alerts when there are errors. These default alerts range from disconnected probes to socket or timeout errors. For example, the HTTP sensors show preconfigured internal alerts according to certain HTTP status codes.
There are also sensors that let you modify the internal sensor alerts. To see if you can modify an alert, check the sensor's settings for customizable options.
See Which HTTP status code leads to which HTTP sensor status? for more information.
See What does error code PExxx mean? for more information on error codes.
Here is an example of a sensor that is in a Down status because of an internal sensor alert.
Another type of alerts are alerts because of limits. These alerts are triggered when the measured value of a sensor is below or above a configured threshold limit. These limits change the sensor's status when they are breached. For example, you can set an SNMP CPU Load sensor to a Warning status whenever it measures values that you consider critical. This sensor then shows up in the alarms list.
This type of alert is only displayed when a value breaches the configured limits. If the value is normal again in the next sensor scan, the sensor returns to an Up status.
- To review or modify threshold limits for single sensors, open the sensor's channel settings via the gear icon in the respective channel gauge or in the channels table.
- Enable alerting based on limits and specify your desired limits in the correct fields.
- You can also optionally add messages that appear in the sensor message on the Overview tab or leave these fields empty to display the default message.
This limit only applies to the respective channel.
For more information on setting limits, see the video tutorial: How to Set Channel Limits
With this configuration, the sensor changes to a Warning status (the gauge needle points to the yellow area in the example below) if the channel values exceed the upper warning limit of 60 percent and the Warning Limit Message is displayed in the sensor status. If the channel value exceeds the upper error limit of 80, the sensor turns to a Down status (the gauge needle points to the red area) and the Error Limit Message is displayed.
You can use Multi-Edit if you want to apply the same limits to sensors of the same type in a batch operation.
- To see all sensors of this type at a glance, just filter for the type: From the main menu bar, select Sensors | By Type | SNMP CPU Load.
- Mark the check boxes of the sensors that you want to edit.
- Click the wrench symbol in the menu.
- Open the Channel Settings tab.
- Select the channel that you want to add a limit for. In this example, we use the channel Total, which is common to all selected sensors.
- Then Enable alerting based on limits and enter the number in the correct field as described earlier.
When you are done, click OK to save these settings—the new limit is then applied to all selected channels from all selected sensors.
Multi-edit is only available for channels that are common to all selected sensors.
Some sensors have predefined limits in the sensor settings, for example the SNMP Linux Disk Free sensor. Refer to the sensor's settings to adjust the limits (or behavior) for these sensors.
Absolute and Delta Values in Limits
The value that needs to be configured in the limits depends on the type of data that the sensor's channel delivers.
For channels that measure absolute values (for CPU load and memory usage, for example), limits are also set using absolute values like in the previous example.
You can also set limits for channels that measure delta (x.xx/sec) values. Let us say you have an SNMP Traffic sensor and want to receive an alert when it reports errors. In this case, you need take into account that this is a delta measurement. In this example, you could set the following limits for the channel Errors in with a standard scanning interval (60 seconds).
- Set the sensor to Warning status when 1 error occurs
- Set the sensor to Down status when 30 errors occur
This example shows how to configure the limits for delta channels.
Because this channel uses per second (delta) measurements, a single error that occurs over a standard 60-second scanning interval is reported as 0.016 # per second. So the warning limit for one single RX error within an interval is 0.1 (errors/sec). To get an alert when there are 30 errors within a scanning interval, the limit needs to be 0.5 (errors/sec).
If no new errors occur in the next scanning interval, the sensor turns back to an Up status. To ensure that you do not miss any notifications for this sensor, set a notification trigger with 0 seconds. For more information, see Notification Triggers as Extensions to Alerts below.
PRTG also uses lookups for some sensors. In general, lookups make data more human friendly because they map status values as returned by a device (usually integers) to more informative expressions in words that show you the status of a monitored device as a clear message.
Additionally, lookups can also define the sensor status that is shown in correlation with certain status codes, just like sensor channel limits can define a sensor status, too. For example, PRTG can show a sensor in a gray Inactive status with channel values, provided by lookups, like Inactive instead of a numerical value like -1.
The various states displayed in gauges always follow the clockwise order Up (green) < Warning (yellow) < Down (red) < Unknown (gray).
You can also modify standard lookups to include new definitions according to your monitoring needs. The following code illustrates the lookup definition for the paper status of the SNMP HP LaserJet Hardware sensor:
<?xml version="1.0" encoding="UTF-8"?>
<ValueLookup id="oid.paessler.hplaserjet.paperstatus" desiredValue="0"
<SingleInt state="Ok" value="0">
<SingleInt state="Error" value="1">
Out of Paper or No Cassette Loaded
<SingleInt state="Error" value="2">
Manual Paper Feed Required
If you do not want to be alerted in case the printer is out of paper, you basically have to do the following:
- Copy the file oid.paessler.hplaserjet.paperstatus.ovl to the \lookups\custom subfolder of the PRTG program directory (ensure you do not change the file name!).
- Open this file with an editor.
- Replace the first SingleInt state "Error" with "Ok".
- Optionally replace the message "Out of Paper or No Cassette Loaded" with "Ok".
- Save the file.
- Reload the custom lookup folder in PRTG.
The channel gauge now displays the new state and the new message.
For more information on how lookups are defined, see section Define Lookups.
Creating a New Lookup
It is also possible to create new lookups to use with any custom or standard sensor that supports lookups. If you want to create a new lookup with the earlier example, you basically follow the same procedure except you save the lookup file in the \lookups\custom subfolder of the PRTG program directory with a new name, for example oid.paessler.hplaserjet.mynewpaperstatus.ovl. Select the newly created lookup file in the sensor's channel settings.
Custom String Lookups
You can also set up custom string lookups, which can exclusively be used with the SNMP Custom String Lookup sensor. This is useful if you are using an object identifier (OID) that returns a known string value stating possible states or values. To inform PRTG of the possible states or values, you have to create a new lookup file that defines them.
For more information, see the Knowledge Base: Monitor and "Lookup" a SNMP String value.
The status or the data of a sensor can trigger notifications. Using this mechanism, you can configure custom external alerting. Which trigger types are available depends on the kind of object you edit. You can define triggers that are activated by an 'on change' event. Some sensors offer the option to send such a trigger whenever sensor values have changed.
Before you set up a change trigger, make sure that the Trigger 'change' notification setting is enabled in the sensor's settings, otherwise the notification is never sent.
Select the Notification Triggers tab to create a change trigger for this sensor. Hover over and select Add Change Trigger from the menu to add a new trigger, or click the Edit button next to an existing trigger to change it. Every trigger provokes one or more notifications.
Save the trigger. You are now notified when sensor values change.
For more information on triggers, see section Sensor Notification Triggers Settings.
After you have set up alerts using limits or lookups, you can complement them with state-based triggers. On the Notification Triggers tab, hover over and select Add State Trigger from the menu to add a new trigger.
Save the trigger. You are now notified according to the notification settings.
If you want to receive alerts for more than one status, you have to add a trigger for each status.
A sensor's status can change for other reasons such as internal sensor alerts. If you configure state-based triggers, you are notified of these changes as well.
Knowledge Base: Monitor and "Lookup" a SNMP String value
Knowledge Base: Which HTTP status code leads to which HTTP sensor status?
Knowledge Base: What does error code PExxx mean?