Event Classifier Reduces Notification Overload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current notification systems for monitoring multi-component systems, such as cloud networks, generate excessive alerts due to immediate notifications for every problem, leading to information overload and potential delays in timely issue detection.
Innovation Solution
A notification system that classifies event signals into predefined event classes based on status and attribute parameters, generating user notifications only for 'bad state' events, with optional aggregation of similar events and customizable notification settings to reduce alert volume while ensuring prompt information delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If immediate notifications are generated for every detected problem in the network, then the system administrator is informed promptly about issues, but the number of notification messages becomes excessively large leading to information overload
Solution Approach 1:
Multiple individual device offline events are merged into a single network-level notification when they share a common cause (e.g., uplink failure). The notification system aggregates events from multiple devices and sends one consolidated alert representing the group, thereby reducing the total number of messages while maintaining timely notification.
Solution Approach 2:
The notification system introduces an intermediary processing layer between event detection and alert generation. This intermediary analyzes the cause of events, determines whether multiple events stem from a common root cause, and selectively aggregates or separates notifications based on this analysis, optimizing the balance between promptness and message volume.
2Quantity of substance
If offline time threshold is introduced before generating alerts, then the number of notifications is reduced, but the system administrator is informed less timely about potential issues
Solution Approach 1:
The notification system dynamically adjusts the effective threshold based on the nature and cause of events. For critical issues like network uplink failures, the system generates immediate notifications without waiting for a timeout period. For less critical or transient issues, the system applies the offline time threshold to filter out false positives. This dynamic behavior allows timely notification for important issues while reducing noise for minor problems.
3Reliability
If each device generates individual alerts when offline, then comprehensive monitoring is achieved, but the notification system becomes complex and difficult to manage
Solution Approach 1:
The notification system segments alerts into different hierarchical levels: device-level individual alerts and network-level aggregated alerts. This segmentation allows the system to maintain comprehensive monitoring at the device level while presenting a simplified, consolidated view at the network level, reducing the apparent complexity for system administrators.
Solution Approach 2:
The notification system implements a universal aggregation mechanism that works across multiple device types and failure modes. A single aggregation process handles various event types (offline events, error events, warnings) from different devices, applying consistent logic to determine whether to aggregate or notify individually. This universal approach simplifies the system compared to having device-specific notification logic.
Data Source
AI summary
A notification system for monitoring a multi-component system is disclosed. The notification system comprises a notification processing module and an event processing module. The event processing module comprises an event input and an event classifier module. The event input is configured to receive an event signal that comprises at least one status parameter being indicative of at least one of a status and a status change of at least one component of the multi-component system. The event classifier module is configured to classify the event signal into at least one event class based on the at least one status parameter. The notification processing module is configured to generate no user notification or at least one user notification based on the at least one event class. Further, a method of monitoring a multi-component system is disclosed.