Correlation Rule weighting

Correlation rule weighting can be used to override existing cause-symptom relationships created by a correlation rule with better correlations available from a more recently matched correlation rule. Better possible correlations are indicated by assigning a higher weighting to the correlation rule. A TBEC rule with a higher weighting overrides an existing cause-symptom relationship previously created by another correlation rule with a lower weighting.

For example, two events are received:

  • an application failure event
  • a database problem event for the database used by the impacted application

The database failure should be used as cause for the application failure. Subsequently, an application server down event for the application server that our application is running on is received. The new event is a more appropriate cause, and should replace the database problem event.

Each correlation rule includes a rule weight: lowest, low, normal, high, or highest.

Even if an event already has a cause assigned to it, another correlation rule can replace the cause if that rule has a higher rule weight than the rule that assigned the event’s current cause.

Note Causes can only be replaced by “better” causes, and cannot be removed.

Event history lines record changes to the assigned cause.

Assigning a new cause to an already closed symptom does not reopen that symptom.

Example

Let us assume that we have the two following correlation rules:

  • TBEC Rule 1

    Database problem on database causes application failure on application if the application uses the database.

    Weight: Low

  • TBEC Rule 2

    Application server down on application server causes application failure on application if the application runs on the application server.

    Weight: Normal

We receive a database problem event on Database-1 and an application failure event on Application-17, which uses Database-1. TBEC Rule 1 matches both events and the database problem event became the cause of the application failure event.

Later, we receive an application server down event on Application Server-42, on which Application-17 is running. TBEC Rule 2 matches the application server down event with the application failure event. The cause of the application failure event is changed to be the application server down event, because TBEC Rule 2 has a higher weighting (Normal) than TBEC Rule 1 (Low).