Skip to content

SIPMon Anomaly Detection

Metric Configuration

The Metric Configuration feature within the SIPMon Anomaly Detection Module enables users to add, manage, and customize performance metrics that drive anomaly detection analysis. By configuring key parameters such as host, service, metric type, and alert severity, users can precisely define which data points to monitor and how anomalies are triggered.

Anomaly Detection HomeFigure: Anomaly Detection Dashboard Overview

Key Features

  • Add Metric: Create new metric entries using an intuitive form interface with guided dropdown selections. Add Metric FormFigure: Add New Metric Interface

  • Edit Metric: Modify existing metric configurations directly from the metric table to adjust detection criteria or alert preferences. Edit Metric FormFigure: Edit Metric Details

  • Delete Metric: Remove unused metrics to clean up the configuration and reduce noise in anomaly detection alerts. Delete MetricFigure: Delete Metric Confirmation

  • Enable/Disable Metric: Activate or deactivate a specific metric without deleting it, offering flexibility for temporary monitoring pauses. Enable/Disable ToggleFigure: Enable/Disable Metric Toggle

  • Contact Management: Assign individual contacts or contact groups to receive alerts based on severity and metric conditions. Contact ManagementFigure: Assigning Contacts for Alerts

  • Alert Severity Levels: Define alert priority as Low, Medium, or High to classify anomaly importance and escalation needs. Alert SeverityFigure: Alert Severity Level Selection

Adding a New Metric

Follow these steps to configure a new metric for anomaly detection:

  1. Initiate Creation: Click the + (Add Metric) button above the metric table. Add Metric ButtonFigure: Add Metric Button

  2. Configure Fields: A popup window will appear. Complete the fields in the following order: Add Metric PopupFigure: New Metric Configuration Form

    • Host: Select the target host from the dropdown list. Host SelectionFigure: Select Target Host

    • Service: Choose the related service linked to the selected host. Service SelectionFigure: Select Related Service

    • Metric: Select the specific performance metric to monitor. Metric SelectionFigure: Choose Performance Metric

    • Alert Severity: Set the alert level (Low, Medium, or High). Severity SelectionFigure: Define Alert Severity

    • Contact(s): Choose one or multiple contacts to notify. Contact SelectionFigure: Select Alert Contacts

    • Contact Group(s): Select one or multiple contact groups for broader notifications. Contact Group SelectionFigure: Select Contact Groups

    • State: Enable or disable the metric as needed instantly. Enable ToggleFigure: Set Initial State

  3. Save: Click Submit to save the configuration or Cancel to discard changes. Submit ButtonFigure: Save Configuration

Managing Existing Metrics

  • Edit: Click the Edit icon in the metric table to modify metric details or alert assignments. Edit IconFigure: Edit Metric Action

  • Delete: Click the Delete icon to permanently remove a metric entry from the system. Delete IconFigure: Delete Metric Action

Quick Anomaly Graphs

The Quick Anomaly Graphs feature provides a fast and visual method to analyze system performance trends and identify anomalies in real time. This functionality generates multiple graph types—including Actual vs Predicted Over Time, Cumulative Trend Over Time, and Anomaly Severity Distribution—for deeper insights.

Quick Anomaly Graph OverviewFigure: Quick Anomaly Graph Dashboard

Features

  • Dynamic Visualization: Quickly visualize anomalies and performance deviations.
  • Multiple Graph Views:
    • Actual vs Predicted Over Time: Compare real-time performance values with predicted baselines.
    • Cumulative Trend Over Time: Observe long-term performance accumulation trends.
    • Anomaly Severity Distribution: View the proportion of detected anomalies by severity level.
  • Customizable Filters: Select specific hosts, services, metrics, and time ranges to refine analysis.
  • Date Range Picker: Define the desired start and end dates for specific period analysis.
  • Instant Generation: Generate graphs dynamically by clicking "Get Anomalies", with real-time results.

Accessing & Generating Graphs

  1. Navigate: Open the Anomaly Detection page and scroll to the Quick Anomaly Graphs section.
  2. Configure Filters: Select the parameters in the following order:
    • Host: Choose the host to analyze.
    • Service: Select the associated service. Service SelectionFigure: Select Service for Graph
    • Metric: Pick the specific metric (e.g., CPU usage, response time). Metric SelectionFigure: Select Metric for Graph
  3. Set Time Range:
    • Select the Start Date. Start DateFigure: Start Date Picker
    • Select the End Date. End DateFigure: End Date Picker
  4. Generate: Click the Get Anomalies button. Get Anomalies ButtonFigure: Generate Graph Action

The selected graphs (Actual vs Predicted, Cumulative Trend, Severity Distribution) will be displayed automatically after processing.

Summary

The Quick Anomaly Graphs feature offers an efficient way to visualize system anomalies without complex configuration. By selecting key parameters coverage, users can instantly generate interactive graphs that highlight deviations, predict trends, and assess severity distributions for informed decision-making.