In the Amsterdam release, VoLTE use case has a requirement on fault correlation and auto-healing, which is closely related to Control Loop, including but not limited to CLAMP, DCAE, Policy, Holmes, Multi-Cloud and VF-C projects.

Story Overview

Example for R1


VNF Fault

EPC S/P-GW – The slave MPU board is offline or abnormal

Retart VNF

VM Fault

The guest OS has indicated a failure, requiring VM restart

I don't know how to merge this column. (sad)

Refer to Holmes rule
public class VesAlarm{
    private String eventId;
    private String sourceName;
    private String specificProblem;
    private long startEpochMicrosec;
    // getters and setters are ommitted for brevity
Refer to Holmes rule
package dcae.ves.test
import org.onap.some.related.packages;

// Alarm specific problems used in the following rules:
// specificProblem
// The slave MPU board is offline or abnormal
// The guest OS has indicated a failure, requiring VM restart

// the entity of the rule
rule "SameVNF_Relation_Rule"
salience 120
no-loop true
        $root : VesAlarm(                                  
            $sourceName: sourceName, sourceName != null && !sourceName.equals(""),
            specificProblem in ( "The guest OS has indicated a failure, requiring VM restart" ),
            $eventId: eventId)                                 
        $child : VesAlarm( eventId != $eventId,
            CorrelationUtil.getInstance().isTopologicallyRelated(sourceName, $sourceName),
            specificProblem in ("The slave MPU board is offline or abnormal"),
            this after [-60s, 60s] $root)                      
  • No labels


  1. Does EMS directly report event to VES collector as shown in the document or through VFC? 

    1. There is an adapter (belongs to VF-C) to convert data format, and then report to VES collector. The process like VNF → EMS → adapter (in VF-C) → VES collector.

      1. Thanks for the clarification

  2. Is the auto scaling functionality available for the VoLTE demo in ONAP Amsterdam?

    1. For Amsterdam release, it only containes auto healing close loop for VoLTE use case .The auto scaling close loop for VoLTE is included in ONAP Beijing planning. 

      1. Thanks.

        It is a limitation of the VoLTE demo or of the ONAP Amsterdam release itself ? Basically I need to understand if it is possible to implement auto healing for a custom VNF in Amsterdam release.


        1. This material describes a self-healing closed-loop scene for VOLTE use case in Amsterdam release. Self-healing action is implemented by VF-C in this use case. If you just want to heal a VNF you can try to invoke VF-C heal interface directly.

          1. Thanks, I'll try that. Is something similar available for scaling as well ?