Anomaly Detection for Single Page Applications
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional performance metric monitoring and root cause analysis techniques are inadequate for single page applications (SPAs) due to their newer loading behaviors, making it challenging to analyze load times and resource performance effectively.
Innovation Solution
A device detects anomalous features in page load information, identifies types of anomalies (page, resource, or domain anomalies) based on resource anomalies, and performs mitigation actions accordingly, using a combination of page load monitoring, anomaly detection algorithms, and machine learning techniques to correlate page load speed with resource performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional performance metric monitoring techniques are used, then monitoring of web services is simple, but they are inadequate for single page applications with newer loading behaviors
Solution Approach 1:
The patent segments the monitoring approach by introducing separate agents (browser agents, server agents, network agents) that each collect specific types of data independently. This segmentation allows the system to adapt to SPA loading behaviors without requiring complete redesign of the monitoring architecture, as each agent can be independently configured and deployed.
Solution Approach 2:
The patent introduces intermediary components including a data collector that aggregates data from multiple agents, and a machine learning model that serves as an intermediary between raw data and anomaly detection. These intermediaries enable the system to handle complex SPA monitoring requirements while maintaining manageable system architecture.
2Loss of information
If comprehensive monitoring data is collected across disparate systems, then more information is available, but it becomes difficult to piece together and connect the chain of events
Solution Approach 1:
The patent merges data from multiple sources (browser agents, server agents, network agents) into a unified data structure through the data collector. This consolidation allows comprehensive information to be gathered while simplifying the integration process by providing a centralized point for data aggregation and normalization.
Solution Approach 2:
The patent implements feedback mechanisms where the machine learning model analyzes collected data and generates predictions about anomalies. This feedback loop allows the system to continuously improve its understanding of the data patterns and automatically adjust monitoring priorities, making the complex data integration process more manageable through intelligent feedback-driven adjustments.
3Measurement precision
If traditional root cause analysis based on resource data is used, then analysis is straightforward, but it cannot effectively analyze load times and resource performance in SPAs
Solution Approach 1:
The patent replaces traditional mechanical/manual root cause analysis methods with machine learning-based automated detection. The machine learning model substitutes for manual analysis by automatically identifying anomalies and their root causes from the collected data, significantly improving measurement precision for SPA performance metrics while reducing the difficulty of detection.
Solution Approach 2:
The patent changes the parameters being monitored from traditional resource-centric metrics to include user experience-centric metrics such as load times, interaction performance, and rendering metrics. This parameter transformation enables effective measurement of SPA-specific performance characteristics that traditional methods cannot capture.
4Extent of automation
If automated anomaly detection is implemented, then root causes can be pinpointed effectively, but the system requires sophisticated algorithms and data processing
Solution Approach 1:
The patent segments the automated detection system into specialized agents with specific functions (browser agents for client-side metrics, server agents for backend metrics, network agents for network performance). This segmentation allows sophisticated automation to be achieved while keeping each individual component relatively simple and manageable.
Solution Approach 2:
The patent introduces intermediary components including a data collector that simplifies data aggregation from multiple agents, and a machine learning model that acts as an intelligent intermediary between raw data and anomaly detection results. These intermediaries reduce the overall system complexity by providing specialized processing layers that handle specific aspects of the detection workflow.
Data Source
AI summary
In one embodiment, a device obtains page load information corresponding to a loaded web application. The device detects, based on the page load information, an anomalous feature of the loaded web application. The device identifies a type of the anomalous feature based on a number of resource anomalies within the loaded web application, wherein the type of the anomalous feature is selected from a group consisting of: a page anomaly; a resource anomaly; and a domain anomaly. The device performs one or more mitigation actions according to the type of the anomalous feature.