Microservice Bottleneck Detection via Response Time Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In microservice systems, identifying the root cause of bottlenecks is challenging due to interdependencies among microservice transactions, making it difficult to detect bottlenecks at an early stage before system health is impacted.
Innovation Solution
A computer-implemented method that executes a bottleneck model training process for each microservice, building a bottleneck indicator model using filtered training data and convergence points for expected response times, and executing a bottleneck identification process to provide alerts when bottlenecks are detected by monitoring data streams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If contemporary bottleneck analysis approaches are used to collect system resource information and generate alerts, then bottleneck detection capability is provided, but the ability to identify root cause and detect early-stage bottlenecks deteriorates due to microservice interdependencies
Solution Approach 1:
The patent segments the bottleneck detection problem by creating separate detection mechanisms for different stages: early-stage detection using response time convergence analysis and late-stage detection using traditional resource monitoring. This segmentation allows each mechanism to specialize in its temporal domain, with the early-stage detector focusing on predictive convergence point analysis and the late-stage detector handling traditional bottleneck identification.
Solution Approach 2:
The patent implements preliminary action by detecting bottlenecks at an early stage before they manifest as traditional resource exhaustion issues. The system monitors response times continuously and identifies convergence points where performance degradation is beginning to occur, allowing preventive intervention before the bottleneck becomes critical and affects overall system health.
2Ease of operation
If microservice transactions are monitored as standalone transactions, then monitoring simplicity is maintained, but the ability to detect interdependency-related bottlenecks deteriorates
Solution Approach 1:
The patent merges multiple monitoring perspectives into a unified approach by combining standalone transaction monitoring with interdependency-aware analysis. The system maintains simple individual transaction monitoring while simultaneously aggregating data across transaction chains to detect interdependency patterns, effectively combining both simple monitoring and complex analysis without requiring users to choose one approach.
Solution Approach 2:
The patent introduces an intermediary layer in the form of a centralized monitoring system that sits between individual microservice transactions and the analysis layer. This intermediary aggregates transaction data, identifies relationships between transactions, and provides both simple monitoring views and complex interdependency analysis, mediating between the need for simplicity and the need for comprehensive detection.
3Loss of energy
If traditional bottleneck analysis is performed after system health is impacted, then resource consumption is minimized, but the ability to address issues proactively deteriorates
Solution Approach 1:
The patent performs preliminary action by continuously monitoring response times and identifying convergence points before resource exhaustion occurs. The system detects early signs of bottleneck formation through response time patterns, enabling proactive intervention that prevents the need for intensive resource analysis later when problems have escalated and system health is compromised.
Solution Approach 2:
The patent implements feedback mechanisms that continuously monitor system response times and adjust detection thresholds dynamically. The system learns from historical data to identify normal patterns and detects deviations that indicate emerging bottlenecks, providing continuous feedback that enables proactive resolution before resource consumption spikes and system health deteriorates.
Data Source
AI summary
Aspects of the invention include a computer-implemented method by executing, via a processor, a bottleneck model training process for microservices in a microservice system, wherein for each of the microservices the bottleneck model training process filters out a subset of training data based at least in part on a current situation setting. Building, via the processor, a bottleneck indicator model for each of the microservices using information from the bottleneck model training process, convergence points for an expected response time. Executing, via the processor, a bottleneck identification process for providing system alerts when a bottleneck is identified, wherein the bottleneck identification process uses analysis to monitor a data stream according to the bottleneck indicator model which uses the convergence points for response times.