Threshold

Provides static thresholds on performance data. A threshold is a value that is compared against the predefined threshold configurations. It is evaluated to see whether it violates a specific restriction. The primary objective of thresholding is to determine any violations and to generate alerts. When the value falls outside the acceptable threshold range, the system generates and stores the event condition and forwards it to the Event Management System.

All the real-time metric data from Timeseries Service is written to a Kafka topic that can be used by Threshold Service. It handles the following operations:
  • Static threshold at metric, resource, and group level
  • Can integrate with IBM® Netcool® Operations Insight® Operations Management. When coupled with output from Streaming Analytics, the Telco Network Cloud Manager - Performance can send real-time alerts of threshold violations to send threshold violations to Tivoli® Netcool/OMNIbus.

Supported types of thresholds

Telco Network Cloud Manager - Performance supports burst or static thresholds. It does threshold evaluations on the performance metric data by using period-based data aggregations against the threshold definitions.

Selected performance metrics can be exported from timeseries database to a preconfigured Kafka topic that is compatible with Predictive Insights.

Thresholds define the status of an attribute based on specific conditions. You can enable threshold evaluation on a selected Resource type. A threshold is violated when the result of the collected metric value is evaluated as exceeding (Over) or dropping (Under) to a specified configured threshold level. The actual evaluation and disposition depends on the threshold type, Over, Under, or Band.

Static thresholding is user-defined static values at specific intervals, which analyze data and generate events when a violation occurs for anomaly detection. Static thresholds are of two types, burst and period.

You can define a static threshold for a metric with Threshold Definition configuration page. If these static thresholds are violated for any performance measure on a device or interface, events are generated at a predefined severity level.

Telco Network Cloud Manager - Performance provides the following types of thresholds:
  • Burst

    Ignores the natural network bursts by evaluating how long in a row the violations occurred per resource. Burst thresholds can be set and reset multiple times.

  • Period

    Involves the concept of accumulated duration for which the threshold is violated. This duration is accumulated per resource until the end of the SLA period. The threshold is violated as soon as a data spike accumulated duration by period exceeds a certain duration of time. If the period thresholds are violated, they are not reset until the end of the period.

  • Baseline
    Provides a view into resource behavior outside the normal range. It uses a profiled deviation function to compare the current daily behavior with the baseline as data flows throughout the system (in real time.) What is normal is defined and what is not normal is flagged.
    Note: Baseline thresholding is available through integration with IBM Operations Analytics - Predictive Insights.
    Baseline thresholds are enabled in Technology Packs follows:
    
name=CPU.Utilization.Percent
description="The percentage of CPU utilization"
units=Percent
aliases=[cpuBusyPoll,JuniperERX CPU Utilization,JuniperChassis CPU Utilization avg last 5min,Huawei CPU Utilization,CiscoDevice System CPU Utilization CPM]
properties={
             resource-types="cpu,card,device",
             baselineEnabled="true"
}

Types of violations

Violations can be exported to IBM Tivoli Netcool/OMNIbus, Kafka, and SNMP Trap.

The list of possible values for alarm severity.
  • Critical
  • Major
  • Minor
  • Warning

System configuration pages

The following configuration pages are available in Telco Network Cloud Manager - Performance or defining and viewing threshold violation alerts:
  • Configuring threshold definitions

    You can define threshold profiles for selected metrics.

  • Configuring alarm rules

    You can define the targets and target groups for routing the alarms.

  • Configuring time schedules

    You can define different time schedules to enable or disable threshold monitoring. You have the flexibility to schedule the alarms for the selected metrics at peak hours, off-peak hours, weekdays, and weekend or at a specific hour in a specific day of a specific month.