Integrated test scene prediction method and system based on multi-dimensional data

Abstract

The invention provides an integrated test scene prediction method and system based on multi-dimensional data, and the method comprises the steps: carrying out the multi-source data collection of an integrated test scene based on a dynamic sampling strategy, obtaining integrated test multi-dimensional data, carrying out the multi-dimensional data preprocessing of the integrated test multi-dimensional data, and obtaining a pre-trained multi-dimensional data analysis model; inputting the pre-processed integrated test multi-dimensional data into the multi-dimensional data analysis model to carry out sub-module data analysis, obtaining a pre-trained intelligent regulation and control model, importing a system diagnosis report and a key index deviation degree report into the intelligent regulation and control model as inputs, generating an intelligent regulation and control decision, and carrying out intelligent regulation and control. By means of the method, system faults, performance bottlenecks and potential problems occurring in the integrated test scene can be accurately predicted, intelligent regulation and control decisions are provided for solving the problems, and accurate system performance evaluation and optimization services can be provided for users.

Description

Technical Field

[0001] The present invention relates to the technical field of system performance optimization, and more particularly, to an integrated test scenario prediction method and system based on multi-dimensional data. Background Art

[0002] The field of system performance evaluation and optimization is undergoing a rapid transformation from traditional manual operation and maintenance to intelligent and automated directions, but the industry as a whole still faces multiple challenges and technical bottlenecks.

[0003] On the one hand, most enterprises rely on the experience judgment and manual troubleshooting of operation and maintenance experts. The fault location takes up to several hours or even days, and is easily affected by subjective factors, such as experience deviation and blind area omission. At the same time, the alarm mechanism based on fixed thresholds is difficult to adapt to dynamic business scenarios, such as sudden traffic increase and elastic resource scaling, resulting in a false alarm rate exceeding 40%, and complex link problems cannot be covered.

[0004] On the other hand, multi-source information such as monitoring data, logs, and traces is scattered in different tools, lacking the ability of unified correlation analysis, resulting in fragmented root cause location.

[0005] Therefore, the existing system performance evaluation and optimization technologies are unable to cope with complex and changing dynamic business scenarios, and are difficult to provide high-quality and stable system performance evaluation and optimization services. Summary of the Invention

[0006] In view of the problems mentioned above, in combination with the first aspect of the present invention, embodiments of the present invention provide an integrated test scenario prediction method based on multi-dimensional data, the method comprising:

[0007] Collecting multi-source data for the integrated test scenario based on a dynamic sampling strategy to obtain integrated test multi-dimensional data;

[0008] Performing multi-dimensional data preprocessing on the integrated test multi-dimensional data, where the integrated test multi-dimensional data includes operation data and monitoring data;

[0009] Obtaining a pre-trained multi-dimensional data analysis model, where the multi-dimensional data analysis model includes a quick location module and a depth analysis module, and inputting the preprocessed integrated test multi-dimensional data into the multi-dimensional data analysis model for sub-module data analysis to obtain a system diagnosis report and a key index deviation report;

[0010] Obtaining a pre-trained intelligent regulation model, and importing the system diagnosis report and the key index deviation report as inputs into the intelligent regulation model to generate an intelligent regulation decision.

[0011] In yet another aspect, embodiments of the present invention further provide an integrated test scenario prediction system based on multi-dimensional data, comprising:

[0012] An acquisition module, which is used to acquire multi-dimensional data for integration testing;

[0013] A processing module, which is used to perform data preprocessing on the multi-dimensional data for integration testing;

[0014] An analysis module, which is used to obtain a pre-trained multi-dimensional data analysis model and perform data analysis on the multi-dimensional data for integration testing based on the multi-dimensional data analysis model;

[0015] A decision-making module, which is used to obtain a pre-trained intelligent regulation model and generate an intelligent regulation decision based on the intelligent regulation model.

[0016] Based on the above aspects, the embodiment of the present application improves the accuracy of system performance evaluation and optimization by integrating 13 types of data sources for multi-dimensional data coverage, automatically adjusts the data acquisition frequency based on the dynamic sampling strategy, reduces the memory usage, constructs the transaction topology relationship through the graph database, automatically identifies the cross-system bottleneck, integrates the performance deficiencies and potential problems that the system urgently needs to process into a specific diagnostic report through in-depth root cause diagnosis, and generates a targeted regulation decision based on the diagnostic report. In summary, the integrated test scenario prediction method improves the efficiency of system performance evaluation and optimization. Brief Description of the Drawings

[0017] Figure 1 is a schematic execution flow diagram of an integrated test scenario prediction method based on multi-dimensional data provided by an embodiment of the present invention.

[0018] Figure 2 is a schematic diagram of an integrated test scenario prediction system based on multi-dimensional data provided by an embodiment of the present invention. Detailed Embodiments

[0019] The present invention will be specifically described below in conjunction with the accompanying drawings of the specification. Figure 1 is a schematic execution flow diagram of an integrated test scenario prediction method based on multi-dimensional data provided by an embodiment of the present invention. The integrated test scenario prediction method based on multi-dimensional data will be introduced in detail below.

[0020] Step S1, perform multi-source data acquisition on the integrated test scenario based on the dynamic sampling strategy to obtain multi-dimensional data for integration testing.

[0021] Specifically, taking the three system load indicators of CPU utilization rate, QPS, and memory occupancy rate as input variables, a Gaussian process regression model is constructed to output the predicted mean and variance of the multi-dimensional data sampling rate. Based on the acquisition function, the optimal sampling rate is selected. When the system is at the baseline load, full data sampling is performed. The baseline load is expressed as the CPU occupancy rate being lower than 60%. When the load increases, the lowest sampling rate that can minimize resource occupancy is selected. The load increase is expressed as the CPU occupancy rate rising above 80%. The system hard limit is used as the optimization boundary condition, and in each iteration, the optimization boundary condition is integrated into the acquisition function through the Lagrange multiplier method to obtain the optimal sampling rate. The system hard limit is expressed as the QPS threshold being 5000 tps.

[0022] Furthermore, 13 types of core data, including integrated test scenario operation data, infrastructure monitoring data, application performance data, database special indicators, network transmission data, end-user experience data, log tracking data, security audit data, configuration change data, business context data, AI training data, third-party system data, and chaos engineering data, are obtained, and the core data is integrated to obtain integrated test multi-dimensional data. For example, full-link call chain data is captured through bytecode injection probes. The full-link call chain data includes three core parameters: transaction ID, response time, and throughput. The Prometheus + Sigar + Oshi architecture is used to collect infrastructure monitoring data. The infrastructure monitoring data includes server CPU, MEM, IO, JVM stack, and middleware thread pool, and the infrastructure monitoring data collection supports dynamic expansion of monitoring plugins. System logs are collected in real time through Netty + Websocket + Logback. The system logs are expressed as DEBUG, INFO, and WARN at the log level, thread ID, stack trace depth, and transaction context ID, and regular expressions are used to parse key events. The user feedback experience data is obtained by docking with the user behavior analysis system through the REST API. The user feedback experience data includes operation heat maps, abnormal interruption records, operation step sequences, interface element IDs, and interruption reason classifications. The interruption reason classifications include system timeouts, UI abnormalities, and business process blockages.

[0023] It can be understood that through cross-analysis of the above 13 types of data, full-stack problem location, intelligent root cause analysis, and dynamic test optimization can be achieved. For example, full-stack problem location can be expressed as accurately identifying resource competition bottlenecks by combining database deadlock logs and thread pool monitoring. Intelligent root cause analysis can be expressed as using an AI model to correlate network packet loss and business timeout events. Dynamic test optimization can be expressed as automatically adjusting the parameters of the stress test model based on historical performance data.

[0024] Step S2, perform multi-dimensional data preprocessing on the integrated test multi-dimensional data, where the integrated test multi-dimensional data includes operation data and monitoring data.

[0025] Specifically, perform multi-dimensional spatio-temporal alignment on the integrated test multi-dimensional data. The multi-dimensional spatio-temporal alignment is manifested as using a sliding window compensation mechanism and the NTP protocol to calibrate the timestamps of multi-source data. For example, use a timestamp correction algorithm to solve the clock deviation problem between multiple distributed performance test devices and sampling probes. The clock deviation problem is expressed as a maximum allowable deviation of ±500 ms. Use the sliding window compensation mechanism to achieve millisecond-level alignment. Perform dimension fusion on the integrated test multi-dimensional data. The dimension fusion is expressed as associating TraceID with logs, monitoring, and business data through the Snowflake algorithm and constructing a multi-dimensional data analysis view. Perform data denoising on the integrated test multi-dimensional data. The data denoising is expressed as applying the 3σ criterion to filter out response time outliers and removing data outliers through the Isolation Forest algorithm. For example, establish an intelligent data denoising system for three-layer data processing. The first layer filters and removes outliers based on rules, such as response time > 3σ. The second layer processes and predicts the data fluctuation range based on the LSTM neural network to identify hidden anomalies. The third layer optimizes and detects data outliers through the Isolation Forest algorithm and uses two-way TLS authentication, such as RSA-2048 keys. The data packet encryption uses the AES-GCM-256 algorithm, and the transport layer enables the KeepAlive keep-alive mechanism, that is, the timeout time is 300 s and the interval time is 60 s.

[0026] Step S3, obtain a pre-trained multi-dimensional data analysis model, input the preprocessed integrated test multi-dimensional data into the multi-dimensional data analysis model for module-by-module data analysis, and obtain a system diagnosis report and a key indicator deviation report.

[0027] In this embodiment, Step S3 includes:

[0028] Step S31, perform anomaly detection based on the quick positioning module in the multi-dimensional data analysis model to obtain a problem candidate set and a key indicator deviation report.

[0029] Specifically, calculate the mean and standard deviation of the integrated test multi-dimensional data, compare the calculated mean and standard deviation with the preset threshold, and mark the data outside the preset threshold range as abnormal data. For example, use the 3σ criterion to identify outliers in CPU usage rate and TPS volatility for test scenario running data, infrastructure monitoring data, and network transmission data, set dynamic thresholds such as CPU > 80% to trigger alarms, construct a UV-TPS curve, use the sliding window algorithm to automatically detect inflection points, locate the time slices of performance mutations, trigger dynamic alarms when abnormal data appears continuously, obtain alarm event data, and cluster similar alarm event data to generate a problem candidate set.

[0030] Take the mean of the integrated test multi-dimensional data in the historical normal period as the reference baseline, compare the deviation value between the current data and the reference baseline, and when the deviation value between the reference baseline and the current data is greater than the preset threshold, mark the current data as a potential risk point and generate a critical metric deviation report.

[0031] Step S32, perform in-depth root cause diagnosis based on the in-depth analysis module in the multi-dimensional data analysis model to obtain a system diagnosis report.

[0032] Specifically, in-depth root cause diagnosis divides the integrated test multi-dimensional data into four dimensions: time series anomaly detection, bottleneck node location, association rule mining, and root cause classification model for data analysis through the linkage of multiple algorithms. In the time series anomaly detection dimension, by using the DTW algorithm, dynamically align the current performance curve with the typical patterns in the obtained historical fault library, calculate the similarity, and obtain the diagnosis result of the abnormal period. For example, 采 Use the LSTM model to predict the trend of JVM heap memory usage, give an early warning of OOM risk 30 minutes in advance, and calculate the correlation between the number of database deadlocks and application response time through the Pearson correlation coefficient; in the bottleneck node location dimension, construct a service call topology graph by using the PageRank algorithm, where the node weight of the service call topology graph is determined by the product of the call frequency and the average time-consuming; obtain the diagnosis result of the key node. In the association rule mining dimension, analyze the full-link event log by using the FP-Growth algorithm, extract frequent item sets, and obtain the diagnosis result of the association rule. For example, divide the high-frequency operation area and the performance-sensitive path by clustering the user operation heat map through K-means, discover the strong association rule presented between the network packet loss rate > 2% and the database response time > 500ms by using the Apriori algorithm, and discover the chain relationship between the decrease in cache hit rate caused by high concurrency and the increase in DB lock waiting by using the FP-Growth algorithm; in the root cause classification model dimension, construct a multi-dimensional feature matrix, classify the bottleneck types through the random forest algorithm, construct a knowledge graph of bottleneck types, and obtain the diagnosis result of the bottleneck type.

[0033] Further, perform dimensional fusion on the four-dimensional diagnostic results to obtain dimensionally fused data, construct a performance knowledge graph based on the dimensionally fused data, and output a system diagnostic report containing abnormal time periods, key nodes, association rules, and bottleneck types based on the performance knowledge graph. For example, use a topology graph to present the cross-system bottleneck link and identify the proportion of resource consumption on the critical path, and use a heat map to display the abnormal time slices and associated metrics.

[0034] Step S4: Obtain a pre-trained intelligent regulation model, import the system diagnostic report and the key indicator deviation report as inputs into the intelligent regulation model, and generate an intelligent regulation decision.

[0035] Specifically, the intelligent regulation model calls the performance knowledge graph to match historical cases, retrieves verified optimization solutions, combines LSTM to predict future indicator trends, generates a hierarchical regulation decision. The hierarchical regulation decision includes emergency regulation and long-term correction. After executing the regulation decision, monitor the recovery effect and extract the current case features, update the performance knowledge graph nodes through incremental learning, and increase the weights of high-success-rate optimization solutions and decrease the weights of low-success-rate optimization solutions. For example, first construct a knowledge network containing entity relationships such as historical performance problems, environmental parameters, and solution plans based on the Neo4j graph database. The node types of the knowledge network cover performance bottleneck types, and the performance bottleneck types include database deadlocks, thread pool exhaustion, environmental characteristics, and solution plans. Encode real-time monitoring metrics such as the abnormal time slice features and TPS fluctuation curves of the current system and the service call link and resource dependency relationship into graph embedding vectors through GNN, and use the attention mechanism to calculate their similarity with the historical case subgraph to achieve triple intelligent matching. The triple intelligent matching includes symptom matching, environment matching, and solution migration. Symptom matching is represented as the similarity of the flame graph time-consuming distribution > 85%, environment matching is represented as the configuration parameter difference < 15%, and solution migration is represented as the success rate of the historical repair plan > 90%. At the same time, dynamically update the newly generated performance problems and their solution paths to the graph through the incremental learning mechanism to form a continuously evolving fault prediction system.

[0036] Figure 2 The figure shows a schematic diagram of an integrated test scenario prediction system based on multi-dimensional data according to some embodiments of the present application that can implement the idea of the present application.

[0037] Specifically, an integrated test scenario prediction system based on multi-dimensional data includes:

[0038] An acquisition module, which is used to acquire multi-dimensional data of integrated tests.

[0039] A processing module, which is used to perform data preprocessing on the multi-dimensional data of integrated tests.

[0040] An analysis module, which is used to obtain a pre-trained multi-dimensional data analysis model and perform data analysis on the integrated test multi-dimensional data based on the multi-dimensional data analysis model.

[0041] A decision-making module, which is used to obtain a pre-trained intelligent regulation model and generate an intelligent regulation decision based on the intelligent regulation model.

[0042] The specific usage and functions of this embodiment will be described below:

[0043] First, multi-source data collection is performed on the integrated test scenario based on a dynamic sampling strategy to reduce the memory usage during data collection and obtain integrated test multi-dimensional data. Then, multi-dimensional data preprocessing is performed on the integrated test multi-dimensional data to eliminate the influence of redundant data. Subsequently, a pre-trained multi-dimensional data analysis model is obtained, and the preprocessed integrated test multi-dimensional data is input into the multi-dimensional data analysis model for sub-module data analysis to obtain a system diagnosis report and a key index deviation report, realizing the rapid positioning and comprehensive analysis of performance problems. Finally, a pre-trained intelligent regulation model is obtained, and the system diagnosis report and the key index deviation report are used as inputs and imported into the intelligent regulation model to generate an intelligent regulation decision. Based on this decision, the system performance can be accurately optimized.

[0044] In addition, an embodiment of the present invention further provides an electronic device, including:

[0045] At least one processor; and a memory communicatively connected to at least one processor; wherein the memory stores instructions executable by at least one processor, and the instructions are executed by at least one processor so that at least one processor can execute the method proposed in Embodiment 1 of the present invention.

[0046] The following specifically introduces each component of the electronic device:

[0047] Among them, the processor is the control center of the electronic device, which can be a single processor or a collective term for multiple processing elements. For example, the processor is one or more central processing units (CPUs), or can be an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement Embodiment 1 of the present invention, such as: one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs).

[0048] Among them, the processor can execute various functions of the electronic device by running or executing software programs stored in the memory and calling data stored in the memory.

[0049] The memory is used to store the software program for implementing the solution of the present invention and is controlled by the processor for execution. The specific implementation manner can refer to the above method embodiments and will not be elaborated here.

[0050] The memory can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only (CD-ROM), or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory can be integrated with the processor or exist independently and be coupled to the processor through the interface circuit of the electronic device. The embodiments of the present invention do not make specific limitations in this regard.

[0051] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0052] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can mean that A exists alone, A and B exist simultaneously, or B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally indicates an "or" relationship between the associated objects before and after, but it may also indicate an "and / or" relationship, which can be specifically understood with reference to the context.

[0053] It should be understood that in the embodiments of the present invention, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0054] The above embodiments are only used to illustrate the technical solutions of the present invention, not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An integrated test scenario prediction method based on multi-dimensional data, characterized in that, The method includes: Performing multi-source data collection on the integration test scenario based on a dynamic sampling strategy to obtain multi-dimensional integration test data; Performing multi-dimensional data preprocessing on the multi-dimensional integration test data, where the multi-dimensional integration test data includes operation data and monitoring data; Obtaining a pre-trained multi-dimensional data analysis model, where the multi-dimensional data analysis model includes a quick positioning module and a deep analysis module, inputting the preprocessed multi-dimensional integration test data into the multi-dimensional data analysis model for module-based data analysis, and obtaining a system diagnosis report and a key indicator deviation report; Obtaining a pre-trained intelligent regulation model, taking the system diagnosis report and the key indicator deviation report as inputs and importing them into the intelligent regulation model to generate an intelligent regulation decision.

2. The integrated test scenario prediction method based on multi-dimensional data according to claim 1, wherein The performing multi-source data collection on the integration test scenario based on a dynamic sampling strategy to obtain multi-dimensional integration test data includes: Obtaining 13 types of core data including integration test scenario operation data, infrastructure monitoring data, application performance data, database special indicators, network transmission data, end-user experience data, log tracking data, security audit data, configuration change data, business context data, AI training data, third-party system data, and chaos engineering data, and integrating the core data to obtain multi-dimensional integration test data.

3. The integrated test scenario prediction method based on multi-dimensional data according to claim 1, characterized in that The dynamic sampling strategy includes: Using the CPU utilization rate, QPS, and memory occupancy rate as input variables to construct a Gaussian process regression model, and outputting the predicted mean and variance of the multi-dimensional data sampling rate; Selecting the optimal sampling rate based on the acquisition function, performing full data sampling when the system is at the baseline load, where the baseline load is represented as the CPU occupancy rate being lower than 60%, and when the load increases, selecting the lowest sampling rate that can minimize resource occupancy, where the load increase is represented as the CPU occupancy rate rising above 80%; Using the system hard limit as the optimization boundary condition, and integrating the optimization boundary condition into the acquisition function through the Lagrange multiplier method in each iteration to obtain the optimal sampling rate, where the system hard limit is represented as the QPS threshold being 5000 tps.

4. A method for predicting an integrated test scenario based on multi-dimensional data according to claim 1, characterized in that The obtaining a pre-trained multi-dimensional data analysis model, inputting the preprocessed multi-dimensional integration test data into the multi-dimensional data analysis model for module-based data analysis, and obtaining a system diagnosis report and a key indicator deviation report includes: Performing anomaly detection based on the quick positioning module in the multi-dimensional data analysis model to obtain a problem candidate set and a key indicator deviation report; Performing in-depth root cause diagnosis based on the deep analysis module in the multi-dimensional data analysis model to obtain a system diagnosis report.

5. The integrated test scenario prediction method based on multi-dimensional data according to claim 4, wherein, The performing anomaly detection based on the quick positioning module in the multi-dimensional data analysis model to obtain a problem candidate set and a key indicator deviation report includes: Calculating the mean and standard deviation of the multi-dimensional integration test data, comparing the calculated mean and standard deviation with a preset threshold, marking the data outside the preset threshold range as abnormal data, triggering a dynamic alarm when abnormal data appears continuously, obtaining alarm event data, and clustering the relevant alarm event data to generate a problem candidate set; Take the average value of the multi-dimensional data of the integration test in the normal historical period as the reference baseline, compare the deviation value between the current data and the reference baseline, and when the deviation value between the reference baseline and the current data is greater than the preset threshold, mark the current data as a potential risk point and generate a key indicator deviation report.

6. The integrated test scenario prediction method based on multi-dimensional data according to claim 4, characterized in that Perform in-depth root cause diagnosis based on the in-depth analysis module in the multi-dimensional data analysis model to obtain a system diagnosis report, including: The in-depth root cause diagnosis divides the multi-dimensional data of the integration test into four dimensions: time series anomaly detection, bottleneck node location, association rule mining, and root cause classification model through the linkage of multiple algorithms for data analysis; In the time series anomaly detection dimension, by using the DTW algorithm, the current performance curve is dynamically aligned with the typical patterns in the obtained historical fault library, the similarity is calculated, and the diagnosis result of the abnormal period is obtained; In the bottleneck node location dimension, a service call topology graph is constructed by using the PageRank algorithm, where the node weight of the service call topology graph is determined by the product of the call frequency and the average time consumption, and the diagnosis result of the key node is obtained; In the association rule mining dimension, the FP-Growth algorithm is used to analyze the full-link event log, extract the frequent item set, and obtain the diagnosis result of the association rule; In the root cause classification model dimension, a multi-dimensional feature matrix is constructed, the bottleneck type is classified by the random forest algorithm, and a knowledge graph of the bottleneck type is constructed to obtain the diagnosis result of the bottleneck type; Fuse the four-dimensional diagnosis results to obtain dimension fusion data, construct a performance knowledge graph based on the dimension fusion data, and output a system diagnosis report including the abnormal period, key nodes, association rules, and bottleneck types based on the performance knowledge graph.

7. A method for predicting an integrated test scenario based on multi-dimensional data according to claim 1, characterized in that Obtain the pre-trained intelligent regulation model, import the system diagnosis report and the key indicator deviation report as inputs into the intelligent regulation model, and generate an intelligent regulation decision, including: The intelligent regulation model calls the performance knowledge graph to match historical cases, retrieves the verified optimization solutions, combines with LSTM to predict the future indicator trend, and generates a hierarchical regulation decision, where the hierarchical regulation decision includes emergency regulation and long-term correction; After executing the regulation decision, monitor the recovery effect and extract the current case characteristics, update the nodes of the performance knowledge graph through incremental learning, and increase the weight of the optimization solutions with high success rates and decrease the weight of the optimization solutions with low success rates.

8. A method for predicting an integrated test scenario based on multi-dimensional data according to claim 1, characterized in that, Perform multi-dimensional data preprocessing on the multi-dimensional data of the integration test, including: Perform multi-dimensional space-time alignment on the multi-dimensional data of the integration test, and the multi-dimensional space-time alignment is manifested as using a sliding window compensation mechanism and the NTP protocol to calibrate the timestamps of multi-source data; Perform dimension fusion on the multi-dimensional data of the integration test, and the dimension fusion is expressed as associating TraceID with logs, monitoring, and business data through the Snowflake algorithm and constructing a multi-dimensional data analysis view; Perform data denoising on the multi-dimensional data of the integration test, and the data denoising is expressed as applying the 3σ criterion to filter out response time outliers and removing data outliers through the Isolation Forest algorithm.

9. An integrated test scenario prediction system based on multi-dimensional data, characterized in that, Including: An acquisition module, which is used to acquire the multi-dimensional data of the integration test; A processing module, which is used to perform data preprocessing on the multi-dimensional data of the integration test; An analysis module, which is used to obtain a pre-trained multi-dimensional data analysis model and perform data analysis on the multi-dimensional data of the integration test based on the multi-dimensional data analysis model; A decision-making module, which is used to obtain a pre-trained intelligent regulation model and generate an intelligent regulation decision based on the intelligent regulation model.

Images