SmartSight Alerts
SmartSight Alerts offer a proactive framework to monitor streaming services and receive alerts for potential issues. This guide explains how to configure and use alerts for service monitoring
Alert Types in SmartSight
SmartSight provides two types of alerts: Manual Alerts and AI Alerts.
Manual Alerts
Manual Alerts allow users to set thresholds for Quality of Experience (QoE) parameters over a specific period. When the threshold is crossed, an alert is triggered. Alerts can be customized as follows:
• Threshold Setting:
Thresholds can be manually set by the user or computed based on average metric values across all users of the service.
• Notification:
Alerts send an email (if configured) and a Slack message (if configured) with detailed information when triggered.
• Dimension Configuration:
• All Dimensions: Fires an alert if the overall QoE value across all dimensions exceeds the threshold.
• Selected Dimensions: Monitors individual thresholds for selected dimensions or combinations, triggering an alert when crossed.
Steps to Configure Manual Alerts
Set the Alert Name (e.g., TestManualCDNAlert).
Choose AI Alert as the Alert Type.
Define the Threshold (e.g., 3).
Select the Dimension (e.g., CDN).
Enter the Alert Calculation Period (e.g., 5 mins).
Set the Anomaly Detection Interval (e.g., 5 mins).
Specify the Minimum Attempts (e.g., 5).
Define the Minimum Users (e.g., 3).
Add the Email recipient for metric notifications and set it to Warning.
Click Save to activate the alert.
.
Key Definitions
• Alert Calculation Period:
Time window (in seconds) considered for calculating the alert.
• Anomaly Detection Interval:
Time interval (in seconds) after which the alert check runs.
Alert Message
This report displays a graph for the QoE parameter, Buffering Ratio.
• Green Band: Represents the alert threshold.
• Red Dots: Indicate alerts triggered by specific dimension values.
Detailed information is provided in the table below.
AI Alerts
AI Alerts function similarly to Manual Alerts but with key differences:
1. No Dimension Pre-selection:
AI Alerts automatically analyze data for anomalies over the specified period without requiring predefined dimension values.
2. Root Cause Analysis:
When an anomaly is detected, AI identifies the dimensions most likely contributing to the issue. It pinpoints the dimension and its value with the greatest impact, providing actionable insights.
Steps to Configure AI Alerts
Set the Alert Name (e.g., TestAICDNAlert).
Choose AI Alert as the Alert Type.
Define the Threshold (e.g., 3).
Enter the Alert Calculation Period (e.g., 5 mins).
Set the Anomaly Detection Interval (e.g., 5 mins).
Specify the Minimum Attempts (e.g., 5).
Define the Minimum Users (e.g., 3).
Add the Email recipient for metric notifications and set it to Warning.
Click Save to activate the alert.
Key Definitions
• Alert Calculation Period:
Time window (in seconds) considered for calculating the alert.
• Anomaly Detection Interval:
Time interval (in seconds) after which the alert check runs.
Alert Message
This report displays a graph for the QoE parameter, Buffering Ratio.
• Green Band: Represents the alert threshold.
• Red Dots: Indicate the top alerts triggeredfor the QoE parameter.
Detailed information is provided in the table below. For each row the dimension which is contributing most to the alert is also mentioned.
Last updated
