云原生系统的性能测试技术研究与实践

Research and practice of performance testing technology for cloud native system

为保障云原生系统的性能质量,设计了分布式的性能测试方案。在并发能力上,以执行机集群的方式突破工具瓶颈,满足云原生系统的高并发需求;在测试效率上,实现自动化性能测试任务,同时测试报告融合监控数据,辅助性能分析。方案包含应用交互层、集成服务层、基础服务层和执行机集群。应用交互层制定典型测试流程,集成服务层实现任务调度和数据整合等,基础服务层和执行机群执行分布式性能测试。工程实践表明,除显著提高测试并发能力外,还可以提高测试、问题诊断的效率,以减少线上环境的性能问题。

In order to guarantee the performance quality of the cloud native system,a distributed performance test system is designed.In terms of system concurrency capability,it breaks through tool performance bottlenecks in the form of executor clusters to meet the high concurrency requirements of cloud native systems.In terms of test efficiency,automated performance testing tasks are implemented,and the test report integrates the monitoring data to assist the performance analysis.The scheme includes application interaction layer,integration service layer,basic service layer and executive machine cluster.The application interaction layer formulates a typical test process.The integration service layer implements test task scheduling and data integration.The basic service layer and execution cluster perform distributed performance testing.Engineering practice shows that in addition to significantly improving the test concurrency capability,it can improve the efficiency of testing and problem diagnosis,and reduce online environmental performance problems.