A method for fault reproduction and diagnosis of multi-tenant SaaS ERP systems

By generating temporary simulated tokens and event joint signatures in a multi-tenant SaaS ERP system, and combining cross-sectional data collection and hierarchical storage from edge computing nodes, the problems of poor isolation and low automation in existing technologies are solved, enabling rapid fault reproduction and root cause localization, and improving the system's diagnostics and operational efficiency.

CN122220139BActive Publication Date: 2026-07-17ZHOUPU DATA TECH NANJING CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
ZHOUPU DATA TECH NANJING CO LTD
Filing Date
2026-05-20
Publication Date
2026-07-17

Smart Images

  • Figure CN122220139B_ABST
    Figure CN122220139B_ABST
Patent Text Reader

Abstract

This invention provides a method for fault reproduction and diagnosis of multi-tenant SaaS ERP systems, belonging to the fields of cloud computing and software testing technology. Through a tenant-level shadow database isolation mechanism and baseline snapshot recovery technology, this invention enables the rapid construction of an independent fault reproduction space consistent with the production environment data state for a designated tenant without affecting production operations or other tenants, thus completely solving the resource contention and security isolation problems in a multi-tenant environment. Employing pre-transaction cross-sectional data collection and pre-modification mirror recording technology, combined with multi-dimensional fault characteristics such as SQL error codes, business state machines, and operation timeouts, it can automatically trace the root cause operation along the cross-sectional causal chain and extract the minimum operation sequence, significantly reducing the cost of manual log analysis and repeated trial and error, and compressing fault location time from hours to minutes.
Need to check novelty before this filing date? Find Prior Art