UIM Alarm created following a group alarm policy instead of the device alarm policy

Products

• DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

UIM generates an alarm using a group alarm policy when a device alarm policy has been created with precedence over the group policy.

See the example below.

Alarm:

INC7470633

2025-06-08 02:20:41

Disk Free (%) on E:/ for XXXXX is at 9.42 %. It has violated threshold of 10.00 percent

The Group Alarm Policy is configured with a threshold of 10% while the Device Alarm Policy is configured for 15%.

Device Alarm Policy configuration:

 

Environment

• UIM 23.4CU3

Cause

In this case, the source had changed from short to fqdn.  QOS metrics had been collected after the source change.  The device alarm policy was created using the obsolete source and the currently collected metrics do not match this source.  The group alarm policy has wildcard matching for the source, so this is the alarm that triggers.

Resolution

This issue can be resolved by removing the obsolete records from the database and recreating the device alarm policy.

• Remove the obsolete information from the database.
  • Open SLM tool in the OC interface
  • Select the Active QoS Objects from the top bar
  • Filter the metrics by probe and source until the obsolete metrics are found
  • Choose Delete->Delete Object(s) from the top menu
  • Deactivate/Activate the data_engine probe.
• Delete the old device alarm policies (if they still exist)
• Create new device alarm policies using the desired settings

You can check in the plugin_metric.cfg file for the robot to confirm that the old policy is removed and the new one is added with the correct source. 

The group policy will have a ~.* (wildcard) source:

   <policy_66>
      policyProfile = 32897
      <cdm>
         <metric_376>
            alarm = true
            qos_name = QOS_DISK_FREE_PERC
            qos_source = ~.*

The device policy will have a specific source:

   <policy_177>
      policyProfile = 60838
      <cdm>
         <metric_460>
            alarm = true
            qos_name = QOS_DISK_FREE_PERC
            qos_source = sampleserver.domain.local

Once this has been completed, the alarms should be created according to the policy when thresholds are breached.