Skip to main content

Anomaly Events

Support
  • Updated

In some circumstances, the machine may work not as well or expected, and this situation can be traced by the use of anomalies.

We can have different levels of anomalies: 

  • Major: the product is still working, but something wrong has happened that affects product health and could potentially cause a failure in the short term; a short term action is needed to avoid this risk.
  • Minor: the product is still working but something wrong has happened that could potentially affect product health in the long term; there is no risk of short term failure, therefore a programmable action is enough to deal with this case.

 An Anomaly is described by:

  • Name: the name of the anomaly (e.g. FAN_SLOWDOWN).

  • Title: the user-friendly name of the event (e.g. Fan running slow).

  • Description: the event description (e.g. The fan seems clogged or damaged).

  • Group: the name of the group to which this event belongs.
    The group can be used to filters and groups events into the DPS pages. 
  • Topics: the set of topics to which this event relates (e.g. CONNECTION, UPTIME, MAINTENANCE).
    Topics are predefined and can be used to filters and groups events into the DPS pages. 
  • Active/Clear Conditions: the metric/property-based conditions used to activate/deactivate the event instance.
    In case of property-based condition, you can use only properties of type DATE.
  • Troubleshooting: a set of remedies the user can follow to fix the problem by himself.

  • Options: a set of variables that can be used to let the user customize the thresholds for event triggering.

  • Alert: how to notify the event to the users.

 

Note that a MAJOR ANOMALY event, by default is associated to CRITICAL severity and ANOMALY category.
A MINOR ANOMALY event, by default is associated to WARNING severity and ANOMALY category. 

 

Creating an Anomaly

To add a new Anomaly event, you should:

  1. Enter the Events / Anomalies page.
  2. Select the Location or Thing Definitions tab, depending on where you want to create the event.
  3. In case of Thing Definition, select the Thing Definition to edit.
  4. Press the Add Event button.
  5. Provide the required information.
  6. Press the Save button and edit the additional information, if needed.

Editing an Anomaly

An Anomaly event is described by the following sections:

General

The main section describing the alert through these properties:

  • Name: specifies the name of the alert, it is a free value (e.g. TEMP WARNING).

  • Title: the alert shown title within the alert list.

  • Description: the text describing the occurred problem.

  • Severity: WARNING or CRITICAL.

  • Category: fixed to ANOMALY

By selecting the Limit the alert visibility depending on the user type checkbox, it is possible to profile the alert visibility to a specific set of User Types.

By selecting the Limit this event computation on a specific time period checkbox, it is possible to define a Start and End day of the year when the event can be activated. 

Within the event Title, and Description it is possible to use placeholders to include information about the thing, the event, and measures.

Dashboard

  • Event Template: the template used to display the event details in a dedicated dashboard.

Active Condition

This section allows defining when the event must be activated.

Note that, an event is evaluated periodically and is triggered only when the active condition changes from false to true and remains true for at least one evaluation interval (e.g. 60 seconds). This means that metric values with a duration of less than one evaluation interval (e.g. 60 seconds) may not be detected in order to correctly trigger the event.

In addition, you can specify the Minimum Active Time the condition must remain true before activating the event. 

 

The condition can be defined by selecting between:

  • Manual: allows specifying whether the event must be manually activated by using the Manual Event Reporting widget. Once the event has been manually activated, it can be cleared automatically by using the Clear Condition or manually by using the Clear action within the Active Alert List widget.
  • Simple: you can select a metric, predicate, and specify a value that can be static (e.g. Temperature 1 > 100 °C), read from another metric (e.g., Temperature 1 > Temperature 2), or read from a property (e.g. Temperature > thing.properties.maxTemperature).
  • Expression: you can select one or more metrics and properties and combine the values into a mathematical expression (e.g. (temp1 + temp2) / 2).
  • Event Based: you can monitor the occurrences of another event over a period of time and activate an anomaly if that event occurs too often. With this condition type, you can select the event to monitor, specify the min occurrences and the relative time frame (e.g. 10 times in 24 hours).
  • Wear Based: it is possible to select a Wear Metric that is checked periodically to see if the wear level has reached the threshold for which a part replacement is requested. No further information needs to be specified for this type of condition; the system automatically uses the max tolerated usage threshold specified in the wear metric.

 

Clear Condition

This section allows defining when the event must be cleared.

If unspecified, the negated active condition is used instead.

The condition can be defined by selecting between:

  • Simple: you can select a metric, predicate, and specify a value that can be static (e.g. Temperature 1 > 100 °C), read from another metric (e.g., Temperature 1 > Temperature 2), or read from a property (e.g. Temperature > thing.properties.maxTemperature).
  • Expression: you can select one or more metrics and properties and combine the values into a mathematical expression (e.g. (temp1 + temp2) / 2).

 

Technical Description and Remedies

Here you can describe more technically the event that has occurred and which are the causes and impacts. Optionally, you can provide a set of remedies, the user can follow to fix the problem by itself.

Remedies are presented to the user through the Thing Troubleshooting widget.

 

Options

This section allows defining options to be used within the event definition and whose value can be redefined by the end-user within the page by suing the thing-options widget.

For more details, see the Options article.

 

Alert

This section allows defining whether the Event must be notified to users through an Alert.

By clicking the Add Alert Button, you can create a new Alert for the event.

 

Event Refactoring

In case you have wrongly defined an anomaly that should instead be a failure, you can convert the event by clicking the Convert to Failure button present in the page bottom.

Related articles

Comments

0 comments

Please sign in to leave a comment.