Suppress Partition Down alerts
search cancel

Suppress Partition Down alerts

book

Article ID: 436996

calendar_today

Updated On:

Products

Network Observability VMware Smart Assurance

Issue/Introduction

  • Users wish to receive individual "Device Down" alerts for every unresponsive device within a partition.
  • The SMARTS IP Availability Manager is grouping unresponsive devices and generating a single "PARTITION DOWN" alarm, which overrides the individual child alerts.
  • By default, Smarts generates a single partition alert when all devices in a partition are unresponsive, instead of independent alarms for each device.

Environment

All supported Smarts releases

Cause

This is the default behavior of the SMARTS IP Availability Manager. When all devices belonging to the same partition are unresponsive, one PARTITION DOWN alarm is generated to reduce alert noise. This event is designed to override individual system down events.

Resolution

To suppress "Partition Down" alerts and ensure that individual "Down" alerts are generated for each device, you must disable the partitioning feature.

Step 1: Disable Partitioning

  1. Set the parameter PartitionEnabled to False in your configuration.
  2. Note: Setting this to False only prevents the creation of future partitions; it does not automatically remove existing ones.

Step 2: Remove Existing Partitions

Because existing partitions are not removed by changing the configuration, you must perform one of the following:

  • Manual Deletion: Manually delete the existing partitions from the topology.
  • Topology Rebuild: Rebuild the topology from scratch to ensure no partitions are defined.
  • Scripted Removal: Use a custom script to remove routers/devices from the partition (e.g., using dmctl operations like remove).

Step 3: Verification

Once partitioning is disabled and existing partitions are removed, the domain will generate individual "Down" alerts for each unresponsive device, even if multiple neighboring devices are down simultaneously.

Important Note: It is strongly recommended to test these changes in a non-production or UAT environment first, as disabling partitioning may significantly increase alert volume in the console.