Auto-scaling Engine Using CMDB Correlation for Root Cause Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Auto-scaling decisions based on monitoring tool alerts can be inaccurate, leading to unnecessary scaling of computing resources and potential business continuity issues, as they may not identify the root cause of issues such as a faulty load balancer rather than the application itself.
Innovation Solution
Integrating Configuration Item (CI) relation data from a Configuration Management Database (CMDB) into the auto-scaling decision-making process to correlate events generated by monitoring tools, ensuring that auto-scaling actions are based on the root cause of problems rather than false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If auto-scaling decisions are made based on monitoring tool alerts, then the system can automatically scale computing resources, but the accuracy of scaling decisions deteriorates due to false positives from infrastructure issues
Solution Approach 1:
The patent introduces an intermediary component that sits between the monitoring tool and the auto-scaling decision-making process. This intermediary analyzes alerts from the monitoring tool, correlates them with configuration item relation data from the CMDB, and determines whether the alert represents a genuine application issue requiring scaling or a false positive caused by infrastructure problems. This intermediary layer filters out false positives while preserving legitimate scaling triggers.
Solution Approach 2:
The system implements a feedback mechanism where scaling decisions are evaluated against actual system state and alert patterns. By continuously monitoring whether scaling actions resolve the underlying issue or merely mask infrastructure problems, the system learns to distinguish between genuine application performance issues and false positives, improving decision accuracy over time.
2Reliability
If scaling actions are taken based on false positive alerts, then computing resources are added to the system, but resource efficiency deteriorates due to unnecessary scaling
Solution Approach 1:
The system performs preliminary analysis of alerts before triggering scaling actions. By checking configuration item relations and infrastructure status in advance, the system determines whether an alert warrants scaling before actually scaling resources. This preliminary filtering prevents unnecessary scaling actions that would waste computing resources while maintaining system availability when scaling is genuinely needed.
3Reliability
If the system scales resources in response to infrastructure problems, then service continuity is maintained, but the root cause resolution deteriorates as infrastructure issues remain unaddressed
Solution Approach 1:
The patent segments the problem analysis into distinct categories: infrastructure issues versus application issues. By analyzing configuration item relations, the system can identify whether an alert originates from infrastructure components (load balancers, networks, hardware) or actual application problems. This segmentation enables differentiated responses: infrastructure issues trigger infrastructure teams for root cause resolution, while application issues trigger appropriate scaling actions, preventing loss of root cause information.
Data Source
AI summary
Provided is a method of auto-scaling. A state change event notification related to a computing resource associated with a computer application is received. Immediate parent configuration items (CIs) associated with the computer application in a configuration management database (CMDB are identified. Status of the immediate parent configuration items (CIs) associated with the computer application is determined. Auto-scaling is performed if the status of the immediate parent configuration items (CIs) associated with the computer application is normal.