CN-122019321-A - Distributed database energized pressure test method and system

CN122019321ACN 122019321 ACN122019321 ACN 122019321ACN-122019321-A

Abstract

The application discloses a distributed database energized pressure test method and a distributed database energized pressure test system, and belongs to the field of aviation retail, wherein the method comprises the steps of collecting heterogeneous audit logs from a production service system, storing the heterogeneous audit logs into a data storage library, carrying out multidimensional analysis on the audit logs, extracting service semantic features, restoring service scenes based on the service semantic features, converting the service scenes into load vectors, applying the load vectors to a tested distributed database system through a pressure test node cluster, collecting performance data of the tested distributed database system under pressure test, carrying out intelligent analysis based on the performance data, and generating test results. The application realizes the full-link, closed-loop and intelligent pressure test of the distributed database cluster in the complex service environment by the multidimensional journal analysis, service scene restoration, abnormal path injection and self-adaptive load scheduling technology.

Inventors

YANG XIN
LIU HAITAO

Assignees

中国民航信息网络股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (12)

1. A distributed database energized pressure test method, the method comprising: collecting heterogeneous audit logs from a production service system and storing the heterogeneous audit logs into a data storage library: Multidimensional analysis is carried out on the audit log, and business semantic features are extracted; Restoring a service scene based on service semantic features, and converting the service scene into a load vector; applying a load vector to the distributed database system under test through the cluster of pressure measurement nodes; and collecting performance data of the tested distributed database system under the pressure test, and performing intelligent analysis based on the performance data to generate a test result.
2. The distributed database energized pressure testing method of claim 1, wherein, The multi-dimensional parsing includes extracting semantic features of an audit log from at least one of transaction, user, and timing dimensions.
3. The distributed database energized pressure testing method of claim 1, further comprising: according to a predefined test plan, the abnormal events in the fault library are integrated into a load vector, and the fault tolerance of the tested distributed database system is tested.
4. The method for distributed database energized pressure testing according to claim 3, wherein, The exception event includes at least one of a network partition, a node failure, or a response delay.
5. The distributed database energized pressure testing method of claim 1, wherein, The pressure testing node cluster is based on a containerized platform, and elastic telescopic scheduling is carried out on the pressure testing node cluster according to the scale of the pressure test.
6. The distributed database energized pressure testing method of claim 1, wherein, The intelligent analysis includes: and determining a business toughness index.
7. The distributed database energized pressure testing method of claim 1, further comprising: generating strategy condition suggestions according to the test result; and carrying out the test again according to the strategy adjustment advice.
8. The distributed database energized pressure testing method of claim 7, wherein, The strategy adjustment proposal comprises adjustment of load intensity or fault type, and self-adaptive exploratory test is realized.
9. A distributed database energized pressure testing system, the system comprising: the collection and storage module is used for collecting heterogeneous audit logs from the production service system and storing the heterogeneous audit logs into the data storage library: the feature taking module is used for carrying out multidimensional analysis on the audit log and extracting business semantic features; The conversion module is used for restoring the service scene based on the service semantic features and converting the service scene into a load vector; The application module is used for applying the load vector to the tested distributed database system through the pressure test node cluster; The testing module is used for collecting performance data of the tested distributed database system under the pressure test, and performing intelligent analysis based on the performance data to generate a testing result.
10. An electronic device, comprising: A processor and a memory; the processor invokes the computer program stored in the memory to perform the distributed database energized pressure test method of any of claims 1 to 8.
11. A computer-readable storage medium comprising, The computer readable storage medium having stored therein a computer program which, when executed by a processor, enables the processor to perform the distributed database energized pressure test method of any of claims 1 to 8.
12. A computer program product, characterized in that, The computer program product comprising a computer program comprising computer program code means stored on a computer readable medium or carrier wave, the computer program code means being configured to cause a computer or processor to control the execution of the distributed database-energized pressure testing method steps according to any of the preceding claims 1 to 8.

Description

Distributed database energized pressure test method and system Technical Field The application belongs to the field of aviation retail, and particularly relates to a distributed database energized pressure test method and system. Background In the air retail industry, core business systems such as orders, payments, inventory, withdrawal and change labels can generate massive audit logs. These logs are critical to troubleshooting and security auditing, and are necessary to ensure stable operation of these systems during peak periods such as "double eleven", holidays, etc., and to perform adequate stress testing. The conventional pressure measuring tool and technology have the following main limitations: 1. The service authenticity is insufficient, the traditional pressure testing tool adopts a script mode to simulate the load, and complex user behaviors, transaction logic and data relativity in a real service scene are difficult to reproduce, so that a significant difference exists between a testing result and production environment performance. 2. The test data is disjointed with the production data, namely, the test data is mostly constructed manually or generated randomly, the scale, distribution and internal relation of the production environment data can not be reflected, and the specific bottleneck (such as specific index failure and hot spot conflict) of the database under the real load is difficult to trigger. 3. Lack of intelligence and insight-the testing process is mostly static and predefined, and the testing strategy cannot be dynamically adjusted according to the testing result. The performance indexes (such as throughput and delay) are mainly concerned, and the comprehensive evaluation of toughness indexes such as fault tolerance capability, self-healing capability, data consistency and the like of the system is lacked. 4. The observability depth is not enough, the testing and monitoring dimension is single, the infrastructure, the database, the application performance and the final business index (such as transaction success rate) cannot be subjected to correlation analysis, so that the root cause is difficult to locate after the problem is found, and the investigation cost is high. These limitations make conventional methods ineffective in testing business continuity in the aviation retail industry under extreme conditions, and a new testing method is needed. Disclosure of Invention In order to solve the above problems, the present application provides a distributed database energized pressure test method, which includes: collecting heterogeneous audit logs from a production service system and storing the heterogeneous audit logs into a data storage library: Multidimensional analysis is carried out on the audit log, and business semantic features are extracted; Restoring a service scene based on service semantic features, and converting the service scene into a load vector; applying a load vector to the distributed database system under test through the cluster of pressure measurement nodes; and collecting performance data of the tested distributed database system under the pressure test, and performing intelligent analysis based on the performance data to generate a test result. According to some embodiments of the disclosure, the multi-dimensional parsing includes extracting semantic features of the audit log from at least one of transaction, user, and timing dimensions. According to some embodiments of the disclosure, the method further comprises: according to a predefined test plan, the abnormal events in the fault library are integrated into a load vector, and the fault tolerance of the tested distributed database system is tested. According to some embodiments of the present disclosure, the exception event includes at least one of a network partition, a node failure, or a response delay. According to some embodiments of the disclosure, the pressure test node cluster is based on a containerized platform, and elastic telescopic scheduling is performed on the pressure test node cluster according to the scale of the pressure test. According to some embodiments of the disclosure, the intelligent analysis comprises: and determining a business toughness index. According to some embodiments of the disclosure, the method further comprises: generating strategy condition suggestions according to the test result; and carrying out the test again according to the strategy adjustment advice. According to some embodiments of the disclosure, the policy adjustment suggestion includes adjusting load strength or fault type to enable adaptive heuristics. The application also provides a distributed database energized pressure test system, which comprises: the collection and storage module is used for collecting heterogeneous audit logs from the production service system and storing the heterogeneous audit logs into the data storage library: the feature taking module is used for carrying out multidimensional