A full-link intelligent fault simulation and evaluation defense method in a microservice scenario

Abstract

A full-link intelligent fault simulation and evaluation defense method in a microservice scenario, comprising the following steps: S1: Encapsulating heterogeneous microservice components without intrusion through sidecar proxy, shielding the differences brought about by different technology stacks; S2: intercepting traffic Through the data traffic proxied by the sidecar, the call link topology of the service is generated, and the monitoring data of the microservice call is obtained; S3: According to the call link topology, fault simulation test cases are automatically generated to test the microservice nodes involved in the call link topology Carry out the test; S4: Automatically execute and generate fault simulation test cases; S5: According to the statistical results of the measured fault impact range of any micro-service node, display it in a chart to evaluate the system's fault tolerance and fault impact range; S6: Realize Service downgrade. The present invention can automatically trigger service degradation, limit the scope of fault influence, and avoid avalanche-like faults in the system.

Description

technical field [0001] The invention relates to the technical field of computer data processing, in particular to a full-link intelligent fault simulation and evaluation defense method in a micro-service scenario. Background technique [0002] With the development and evolution of distributed technology, microservice technology has become an inevitable choice for large-scale distributed IT architecture. In the context of microservice technology, a single user request may pass through multiple service nodes deployed on multiple hosts, forming complex call links, making it difficult to assess the impact of microservice node failures. The current mainstream technology is to intercept the service method through bytecode injection to obtain the call link relationship of the node, but this technology is only applicable to microservices with the same development language. There are also methods that directly develop and modify microservice components to generate call links. The di...

AI Technical Summary

Problems solved by technology

There are also direct development and modification of microservice components to generate call links. The disadvantage of this technical solution is that all monitored microservice components must be customized and modified, which is highly intrusive and relatively expensive.

In the actual production process, the microservice development technology stack needs to be diverse. In order to be compatible with these heterogeneous service components, the use of bytecode injection technology and custom adaptation solutions is limited, that is, existing technologies cannot be compatible Under the premise of a variety of development technology stacks, it can also generate microservice call links at low cost and without intrusion

[0003] After obtaining the microservice call link, in order to confirm the actual impact range of the microservice node failure, it is necessary to automatically simulate the microservice node failure and collect the actual call request flow and response time of the microservice system after the failure occurs, and then Assess the scope of the impact of the failure. The existing public link monitoring system does not have the ability to automatically simulate node failures, and thus does not have the ability to assess the impact of microservice node failures.

[0004] To sum up, the existing technical solutions have poor support for the generation of invocation links in the microservice business system composed of heterogeneous service components. The microservice system needs to be instrumented to monitor and obtain the microservice invocation links, and the access cost is high. Strong intrusion; it is not easy to evaluate the impact range of microservice node failures. It is necessary to manually operate the microservice nodes to simulate the occurrence of failures, and manually correlate the monitoring data corresponding to the failure simulation operation to obtain statistics on the impact of node failures on the system and evaluate microservice node failures. The complexity of the impact degree and scope is high; for faults with a large impact, manual downgrade operations are required to perform service downgrades, which is cumbersome and untimely

Images