Search

CN-121095591-B - Abnormal root cause positioning method and device based on order service system and electronic equipment

CN121095591BCN 121095591 BCN121095591 BCN 121095591BCN-121095591-B

Abstract

The embodiment of the invention discloses an abnormal root cause positioning method, device and electronic equipment based on an order service system. The method comprises the steps of obtaining a multi-dimensional service data set, carrying out image format conversion on the multi-dimensional service data set to obtain a service image data set, carrying out space-time reconstruction anomaly identification and anomaly prediction identification on the multi-dimensional service data set to obtain a service reconstruction anomaly identification information set and a service anomaly prediction information set, generating the service anomaly data set, carrying out index association analysis on the service image data set to obtain an anomaly association index data set, generating a service anomaly causal dependency graph, carrying out anomaly root cause positioning on the service anomaly data set to obtain a service anomaly root cause information set, generating a service anomaly root cause report, and carrying out dynamic adjustment on an order service system. The embodiment can quickly find out abnormal information of the order service system, improve the speed of root cause positioning and improve the stability and performance of the order service system.

Inventors

  • ZHANG TAO
  • CAO QIANG

Assignees

  • 朴道征信有限公司

Dates

Publication Date
20260512
Application Date
20250915

Claims (8)

  1. 1. An abnormal root cause positioning method based on an order service system comprises the following steps: acquiring a multidimensional service data set of an order service system, wherein the multidimensional service data comprises at least one of order data, memory data and bandwidth data; performing image format conversion on the multi-dimensional service data set to obtain a service image data set; The method comprises the steps of carrying out space-time reconstruction anomaly identification on a multi-dimensional service data set to obtain a service reconstruction anomaly identification information set, inputting the multi-dimensional service data set into a first convolution extraction network included in a reconstruction space feature extraction network to obtain a first service space feature vector set, wherein the reconstruction space feature extraction network further comprises a second convolution extraction network and a full-connection layer, inputting the first service space feature vector set into the second convolution extraction network to obtain a second service space feature vector set, inputting the second service space feature vector set into the full-connection layer to obtain a third service space feature vector set, inputting the multi-dimensional service data set into a forward long-short-term memory neural network included in the reconstruction space feature extraction network to obtain a first service time feature vector set, the reconstruction space feature extraction network further comprises a reverse long-short-term memory neural network, inputting the first service time feature vector set into the reverse long-term memory neural network to obtain a second service time feature vector set, utilizing attention to obtain a third service space feature vector set, inputting the time-sequence feature vector set into a time-sequence feature encoder, and the time-sequence feature encoder decoding the time-space feature vector set comprises the time-space-time feature vector set, and the time-sequence feature encoder decoding the time-sequence feature vector set, and the time-sequence feature-sequence-feature-encoding device is obtained by the time-sequence-feature-encoded by the time-sequence-feature-encoded data set, as a service reconfiguration anomaly identification information set; carrying out abnormal prediction recognition on the service image data set to obtain a service abnormal prediction information set; generating a business anomaly data set for the multi-dimensional business data set according to the business reconstruction anomaly identification information set and the business anomaly prediction information set; performing index association analysis on the multidimensional service data set to obtain an abnormal association index data set with association relation with the service abnormal data set; generating a business anomaly causal dependency graph of the multidimensional business data set according to the anomaly associated index data set; Performing second-order anomaly root cause positioning on the service anomaly data set according to the service anomaly causal dependency graph to obtain a service anomaly root cause information set; And inputting the business anomaly data set and the business anomaly root cause information set into an anomaly root cause report generation model to obtain a business anomaly root cause report, and dynamically adjusting the order business system according to the business anomaly root cause report.
  2. 2. The method of claim 1, wherein the performing index association analysis on the multi-dimensional service data set to obtain an abnormal association index data set having an association relationship with the service abnormal data set comprises: According to the business abnormal data set and the multidimensional business data set, randomly generating a business association rule set; vector encoding is carried out on the service association rule set to obtain a service association rule vector set which is used as an initial service rule population; generating a population fitness function set aiming at the initial business rule population, wherein the population fitness function set comprises a population confidence function, a population intelligibility function, a population attribute amplitude function and a population maximum information coefficient function; inputting the initial business rule population into the population fitness function set to obtain a business individual initial fitness group set; Generating an initial population external archive and a screened target initial business rule individual according to the business individual initial fitness group set; based on the initial traffic rule population, the following determination steps are performed: according to the executed times of the determining step, a preset execution times threshold value, the screened target initial business rule individuals and the external archiving adjustment factors, stage updating is carried out on the initial business rule population to obtain an updated business rule population; inputting the updated business rule population into the population fitness function set to obtain a business individual updating fitness group set; Updating the adaptation degree group set according to the business individuals, and updating the external archive of the initial population to obtain the updated external archive of the population; in response to determining that the number of the updated business rule individuals included in the updated population external archive is greater than or equal to a preset archive threshold, removing the individuals from the updated population external archive to obtain a removed external archive; screening the removed external archives to obtain a target screening business rule individual; According to the updated business rule population and the external archive of the initial population, updating the external archive adjustment factor to obtain an updated external archive adjustment factor; And in response to determining that the executed times is greater than or equal to the preset execution times threshold, determining a plurality of multidimensional service data corresponding to a plurality of updated service rule individuals included in the updated population external archive as an abnormal association index data set.
  3. 3. The method of claim 2, wherein after the performing the following determining step based on the initial business rule population, further comprising: In response to determining that the executed times is less than the preset executed times threshold, determining removed external archives, updated business rule populations, target screening business rule individuals, and updated external archive adjustment factors as initial population external archives, initial business rule populations, screened target initial business rule individuals, and external archive adjustment factors, respectively, and determining the sum of the executed times and preset values as executed times to execute the determining step again.
  4. 4. The method of claim 2, wherein the step of updating the initial business rule population in stages according to the number of times the determining step has been performed, a preset number of times of execution threshold, the screened target initial business rule individual, and the external archive adjustment factor, to obtain an updated business rule population, comprises: For each initial business rule individual included in the initial business rule population, performing the following updating steps: Determining an individual screening probability value and an updating stage balance factor according to the executed times and the preset execution times threshold; Responding to the determination that the initial business rule individuals are individuals in a target initial business rule population, and carrying out first position updating on the initial business rule individuals according to the external archiving adjustment factors and the executed times to obtain first updated business rule individuals, wherein the target initial business rule population is a population generated according to the external archiving adjustment factors and the initial business rule population; Responding to the determination that the initial business rule individuals are individuals in a residual initial business rule population, and carrying out second position updating on the initial business rule individuals according to the executed times, the preset execution times threshold and the screened target initial business rule individuals to obtain second updated business rule individuals, wherein the residual initial business rule population is a population obtained by removing the target initial business rule population from the initial business rule population; Performing position correction on the first updated business rule individual or the second updated business rule individual to obtain a first corrected business rule individual serving as an updated business rule individual; Responding to the determination that the individual screening probability value is greater than or equal to the updating stage balance factor, and carrying out third position updating on the initial business rule individual according to the individual screening probability value, the executed times and the preset execution times threshold value to obtain a third updated business rule individual; and carrying out position correction on the third updated business rule individual to obtain a second corrected business rule individual serving as an updated business rule individual.
  5. 5. The method of claim 1, wherein the generating a business anomaly causal dependency graph of the multi-dimensional business dataset from the anomaly association index dataset comprises: Acquiring a historical business anomaly root cause information set which is the same as the anomaly type of the business anomaly data set; Carrying out association fusion on the historical business anomaly root cause information set and the anomaly associated index data set to obtain a target business anomaly index data set; carrying out causal relationship analysis on the target business anomaly index data set to obtain a business anomaly causal data set; Determining a business processing flow component set corresponding to the abnormal associated index data set; carrying out association map construction on the business anomaly causal data set and the business processing flow assembly set to obtain an anomaly index association causal map; And adding the business processing flow assembly set into the abnormal index association causal graph to obtain a business abnormal causal dependency graph.
  6. 6. An abnormal root cause positioning device based on an order service system, comprising: An acquisition unit configured to acquire a multidimensional service data set of the order service system, wherein the multidimensional service data includes at least one of order data, memory data, and bandwidth data; the image format conversion unit is configured to perform image format conversion on the multi-dimensional service data set to obtain a service image data set; The time-space reconstruction anomaly identification unit is configured to perform time-space reconstruction anomaly identification on the multi-dimensional service data set to obtain a service reconstruction anomaly identification information set, and comprises the steps of inputting the multi-dimensional service data set into a first convolution extraction network included in a reconstruction space feature extraction network to obtain a first service space feature vector set, wherein the reconstruction space feature extraction network further comprises a second convolution extraction network and a full-connection layer, inputting the first service space feature vector set into the second convolution extraction network to obtain a second service space feature vector set, inputting the second service space feature vector set into the full-connection layer to obtain a third service space feature vector set, inputting the multi-dimensional service data set into a forward short-time period memory neural network included in the reconstruction space feature extraction network to obtain a first service time feature vector set, the reconstruction space feature extraction network further comprises a reverse short-time period memory neural network, inputting the first service time feature vector set into the reverse short-time period memory neural network to obtain a second service time feature vector set, decoding the service space feature vector set by using a time-space feature vector encoder, and a time-space feature decoder, wherein the time-space feature vector is obtained by using the time-space feature vector encoder, the time-sequence feature encoding device is further integrated with the time-space feature vector encoding function, as a service reconfiguration anomaly identification information set; The abnormal prediction recognition unit is configured to perform abnormal prediction recognition on the business image data set to obtain a business abnormal prediction information set; A first generation unit configured to generate a traffic anomaly data set for the multi-dimensional traffic data set from the traffic reconstruction anomaly identification information set and the traffic anomaly prediction information set; The index association analysis unit is configured to perform index association analysis on the multidimensional service data set to obtain an abnormal association index data set with association relation with the service abnormal data set; a second generation unit configured to generate a business anomaly causal dependency graph of the multi-dimensional business dataset according to the anomaly associated index dataset; the abnormal root cause positioning unit is configured to perform second-order abnormal root cause positioning on the business abnormal data set according to the business abnormal causal dependency graph to obtain a business abnormal root cause information set; the dynamic adjustment unit is configured to input the business anomaly data set and the business anomaly root cause information set into an anomaly root cause report generation model to obtain a business anomaly root cause report, and dynamically adjust the order business system according to the business anomaly root cause report.
  7. 7. An electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon, When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-5.
  8. 8. A computer readable medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method of any of claims 1-5.

Description

Abnormal root cause positioning method and device based on order service system and electronic equipment Technical Field The embodiment of the disclosure relates to the technical field of computers, in particular to an abnormal root cause positioning method, device and electronic equipment based on an order service system. Background With the development of computer technology, various systems are continuously emerging, and one system usually carries various data processing tasks, so that the system is subjected to abnormality detection, and the root cause of the abnormality is determined to be more and more focused, so that the stability of the system is improved, and the system loss is reduced. For the abnormal root cause positioning of the order service system, a mode of using a variation self-encoder to reconstruct and detect the obtained system data set to obtain a system abnormal information set is generally adopted. Then, a dependency graph of the system data is generated. And finally, carrying out root cause analysis on the dependency graph by utilizing a random walk algorithm to obtain an abnormal root cause information set. However, in practice, it is found that when the order service system is subjected to abnormal root cause positioning in the above manner, firstly, because the system data set is high-dimensional time sequence data associated with various indexes, the abnormal root cause positioning accuracy is low because the system data set is subjected to reconstruction abnormal detection only by a variation self-encoder, the internal association relationship of the data cannot be detected, so that the abnormal detection accuracy is low, and the dependency relationship between the system data is only extracted by the dependency relationship graph. The above information disclosed in this background section is only for enhancement of understanding of the background of the disclosed concept and, therefore, it may contain information that does not form the prior art that is known to those of ordinary skill in the art in this country. Disclosure of Invention The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Some embodiments of the present disclosure propose an order service system based abnormal root cause positioning method, apparatus and electronic device to solve one or more of the technical problems mentioned in the background section above. According to the first aspect, some embodiments of the present disclosure provide an anomaly root cause positioning method based on an order service system, which includes obtaining a multidimensional service data set of the order service system, wherein the multidimensional service data set includes at least one of order data, memory data and bandwidth data, performing image format conversion on the multidimensional service data set to obtain a service image data set, performing space-time reconstruction anomaly identification on the multidimensional service data set to obtain a service reconstruction anomaly identification information set, performing anomaly prediction identification on the service image data set to obtain a service anomaly prediction information set, generating a service anomaly data set for the multidimensional service data set according to the service reconstruction anomaly identification information set and the service anomaly prediction information set, performing index correlation analysis on the multidimensional service data set to obtain an anomaly correlation index data set with the service anomaly data set, generating a service anomaly causal dependency graph of the multidimensional service data set according to the anomaly correlation index data set, performing second-order anomaly root cause positioning on the service anomaly data set according to the service anomaly image data set to obtain a service root cause report, and inputting the service root cause report to the anomaly root cause and the anomaly root cause adjustment system according to the service anomaly cause. In a second aspect, some embodiments of the present disclosure provide an anomaly root cause positioning device based on an order service system, comprising an acquisition unit configured to acquire a multidimensional service data set of the order service system, wherein the multidimensional service data includes at least one of order data, memory data, and bandwidth data, an image format conversion unit configured to perform image format conversion on the multidimensional service data set to obtain a service image data set, a space-time reconstruction anomaly identification unit configured to perform space-time reconstruction anomaly identification on the multidimensional service dat