Cloud Installation Anomaly Detection Using Resource Dependency Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud application installation failures are often undetected until the end of the long wait time, causing unnecessary delays and requiring users to manually trace multiple intermediate failures to identify the root cause, which is inefficient and time-consuming.
Innovation Solution
A resource dependency graph is created to define dependencies between resources in a cloud native application, using timelines to detect anomalies and reduce false alarms, and a cause analysis resource is generated to provide better interpretability for troubleshooting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual tracing of intermediate failures is used to identify root cause, then accurate root cause identification is achieved, but user time and effort are excessively consumed
Solution Approach 1:
The patent introduces an intermediary system (anomaly detection system with resource dependency graph) that automatically analyzes intermediate failures and identifies root causes. This mediator processes the complex tracing task that would otherwise require manual user effort, maintaining accurate root cause identification while dramatically reducing user time and effort consumption.
Solution Approach 2:
The system performs preliminary analysis of intermediate failures automatically during the installation process. By pre-computing the resource dependency graph and continuously monitoring resource conditions, the system prepares root cause analysis in advance, eliminating the need for users to manually trace failures after installation completes.
2Loss of time
If anomaly detection is implemented during installation, then user wait time is reduced, but system complexity increases
Solution Approach 1:
The patent segments the anomaly detection function into distinct modular components: resource dependency graph creation, resource condition determination, and anomaly detection logic. This segmentation allows the complex detection system to be built from manageable modules, reducing the perceived system complexity while enabling continuous anomaly monitoring during installation to reduce user wait time.
Solution Approach 2:
The resource dependency graph serves as an intermediary data structure that simplifies the complexity of tracking relationships between installation resources. By pre-defining dependencies in a structured graph format, the system manages complexity internally while providing simple anomaly detection queries to users, reducing wait time without exposing system complexity.
3Measurement precision
If resource dependency graph with time ranges is created, then anomaly detection precision is improved, but data processing complexity increases
Solution Approach 1:
The patent enhances anomaly detection precision by adding time range parameters to resource nodes in the dependency graph. Each resource is associated with expected creation and readiness time ranges, allowing precise anomaly detection when resources fail to meet timing expectations. This parameter enrichment improves detection precision while maintaining manageable complexity through standardized time-based checks.
Solution Approach 2:
The system replaces complex manual analysis mechanics with automated computational checks of time range parameters. Instead of requiring sophisticated algorithms to detect anomalies, the system uses straightforward comparisons of actual versus expected time ranges, improving detection precision through simple, efficient parameter-based logic that reduces processing complexity.
Data Source
AI summary
A method, system, and computer program product that are configured to: create a resource dependency graph that defines dependencies between resources created in an installation of a cloud native application, wherein the resource dependency graph includes a respective time range of each respective one of the resources; for each respective one of the resources, determine a condition of the respective one of the resources based on the respective time range of the respective one of the resources, a success condition of the respective one of the resources, and one or more failure conditions of the respective one of the resources; and create a cause analysis resource that includes the determined condition of each respective one of the resources.