Event Pattern Prediction With Placeholder Grouping for Early IT Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing event management systems fail to detect incomplete event patterns, leading to delayed recognition and response to potential problems in telecommunication and IT environments, thereby increasing mean time to repair and mean time between failures.
Innovation Solution
An event management solution module that provides historical pattern data, detects event patterns, predicts the arrival of remaining events within a predicted period, and alerts operators, as well as forms event groups with placeholder events that are replaced as new events arrive, providing early warning and preparation time for impending issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If event management systems wait for full pattern detection before alerting operators, then false alerts are reduced, but response time increases and incomplete patterns are missed
Solution Approach 1:
The system performs preliminary actions by creating placeholder events for predicted remaining events before they actually occur. When a partial pattern is detected, the system calculates the probability of completion and creates placeholders in advance, allowing operators to prepare responses before the full pattern materializes, thus reducing mean time to repair while maintaining detection accuracy through probability thresholds
Solution Approach 2:
The event pattern detection process is segmented into multiple stages: initial pattern detection, probability calculation, placeholder creation, and completion verification. This segmentation allows the system to alert operators at intermediate stages with high-probability predictions rather than waiting for complete pattern formation, balancing accuracy with response time
2Reliability
If event grouping is only performed for complete patterns, then alert accuracy is maintained, but operator awareness of developing issues is delayed
Solution Approach 1:
Placeholder events serve as intermediaries between partial pattern detection and complete pattern formation. These placeholders represent predicted remaining events and provide operators with early warning information about developing issues. The placeholders maintain alert accuracy through probability thresholds while preserving early warning information by grouping predicted events with currently detected events
Solution Approach 2:
The system performs preliminary grouping actions by creating event groups that include both detected events and predicted placeholder events. This preliminary grouping provides operators with advance awareness of developing issue patterns before they fully materialize, losing no early warning information while maintaining reliability through probability-based filtering
Data Source
AI summary
Predicting the formation of event groups in complex and enterprise-related information technology (IT) environments is provided. In some instances, an event pattern is predicted based upon events arriving within a duration that matches a historical pattern for an event and surpasses a user-defined threshold of events that are received. This predicted event pattern is used to monitor incoming events in order to alert an operator or site reliability engineer (SRE) that a certain percentage of events from the predicted pattern are received. In another instance, receiving an event automatically triggers the formation of an event group, with the event group having information relating to the received event and a series of placeholder events. As more events are received, these newly received events individually replace the placeholder events in the event group.