Service Instance Scheduling Using Multi-Upstream Health Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing service instance scheduling methods cause overall performance decline due to unreasonable fusion of service instances, which is not addressed by current solutions that lack a global perspective on service instance abnormalities.
Innovation Solution
A service instance scheduling method that obtains microservice access data in a bypass mode through a service mesh, evaluates the health degree of service instances using an evaluation indicator, and fuses or recovers instances based on this evaluation to ensure rational scheduling and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If service instance fusion is performed based on limited access data from single upstream service, then local service stability is improved, but overall service performance deteriorates due to false fusion and lack of global perspective
Solution Approach 1:
The patent segments the service instance evaluation process into multiple independent evaluation dimensions (response time, error rate, load, etc.) and collects data from multiple upstream services separately. Each dimension is evaluated independently and then aggregated to form a comprehensive health degree score, avoiding premature fusion decisions based on limited data.
Solution Approach 2:
The patent creates a universal evaluation mechanism that collects access data from multiple upstream services (at least two) to assess the same downstream service instance. This multi-functional approach allows the system to evaluate service instance health from multiple perspectives simultaneously, ensuring comprehensive assessment before fusion decisions are made.
2Speed
If service instance fusion is performed without comprehensive evaluation, then response time is reduced, but service performance deteriorates due to unreasonable fusion scheduling
Solution Approach 1:
The patent performs preliminary evaluation actions by collecting access data from multiple upstream services and calculating comprehensive health degree scores before any fusion decision is made. This preliminary assessment ensures that only service instances with confirmed abnormalities are fused, preventing premature or incorrect fusion that would degrade overall performance.
Solution Approach 2:
The patent implements a feedback mechanism where access data from multiple upstream services continuously flows into the evaluation system. The system calculates health degree scores based on this feedback and adjusts fusion decisions dynamically. This closed-loop feedback ensures that fusion actions are based on accurate, up-to-date information rather than assumptions.
Data Source
AI summary
Embodiments of the present disclosure provide a service instance scheduling method and apparatus, an electronic device, and a storage medium. A microservice access data for a target service is obtained in a bypass mode based on a service mesh, where the microservice access data represents access records of at least two upstream services of the target service with respect to a service instance of the target service. An evaluation indicator of at least one service instance corresponding to the target service is obtained according to the microservice access data, where the evaluation indicator represents a health degree of the service instance. The service instance is fused according to the evaluation indicator of the service instance.