Platform operation performance evaluation method and system based on collaborative intelligent analysis

Abstract

A platform operation performance evaluation method and system based on collaborative intelligent analysis relates to the field of electrical digital data processing, and the method comprises the following steps: establishing a directed data dependency graph according to a data association relationship, and arranging operation performance data according to a time sequence to form a performance fluctuation sequence; calculating association importance of a monitored object in the directed data dependency graph, analyzing a performance fluctuation sequence, and calculating a propagation consistency factor of performance data exceeding a preset performance threshold and a time closeness factor of associated performance data in the performance fluctuation sequence; calculating a propagation weight value between any two monitoring objects, constructing a performance fluctuation propagation sequence of the monitoring objects based on the propagation weight value, and calculating the abnormal influence degree of the monitoring objects according to the performance fluctuation propagation sequence and the directed data dependency graph, and determining the monitoring object with the abnormal influence degree exceeding a preset influence degree threshold value as a target abnormal object. By implementing the method, the accuracy of evaluating the operation performance abnormity of the platform can be improved.

Description

Technical Field

[0001] The present application relates to the field of electronic digital data processing, and in particular to a platform operation performance evaluation method and system based on collaborative intelligent analysis. Background Art

[0002] With the widespread adoption of distributed systems and microservices architectures, platform scale continues to expand, and the interactions between various system components are becoming increasingly complex, placing higher demands on platform performance evaluation and anomaly diagnosis. Accurately identifying the root causes of performance anomalies and analyzing their propagation paths are crucial to ensuring stable platform operation.

[0003] In related technologies, performance monitoring probes are set up to collect performance data from each monitored object and perform anomaly detection based on preset performance thresholds. When a performance anomaly is detected, the possible source of the anomaly is located by analyzing the performance data at the time of the anomaly and combining it with the call relationships between the monitored objects.

[0004] However, due to the complex data dependencies between the various monitored objects in the platform, performance anomalies may spread between objects. The relevant technologies are based on performance data at a single moment for analysis, which cannot effectively track the propagation process of anomalies between objects. It is easy to misjudge or miss the real source of the anomaly, resulting in a decrease in the accuracy of the platform's operating performance anomaly assessment. Summary of the Invention

[0005] The present application provides a platform operation performance evaluation method and system based on collaborative intelligent analysis, which is used to improve the accuracy of platform operation performance abnormality evaluation.

[0006] In the first aspect, the present application provides a platform operation performance evaluation method based on collaborative intelligent analysis, which is applied to the platform operation performance evaluation system. The method includes: obtaining the data association relationship between the monitoring objects in the platform to be evaluated, and establishing a directed data dependency graph based on the data association relationship, wherein each edge in the directed data dependency graph represents the data interaction relationship between the monitoring objects; collecting the operation performance data of the monitoring objects, and arranging the operation performance data in chronological order to form a performance fluctuation sequence; calculating the association importance of the monitoring objects in the directed data dependency graph, and performing weighted calculation based on the number of direct associations and indirect associations of the monitoring objects; analyzing the performance fluctuation sequence, and calculating the propagation consistency factor of the performance data exceeding the preset performance threshold and the time closeness of the associated performance data in the performance fluctuation sequence. Factor, the propagation consistency factor represents the degree of propagation of performance data between associated objects, and the time closeness factor represents the degree of temporal correlation of associated performance data; the propagation weight value between any two monitored objects is calculated and a performance fluctuation propagation sequence of the monitored objects is constructed based on the propagation weight value. The propagation weight value is obtained by weighted calculation of the associated importance, the propagation consistency factor and the time closeness factor. The propagation sequence represents the transmission order of performance anomalies between each monitored object; according to the performance fluctuation propagation sequence and the directed data dependency graph, the abnormal impact of the monitored object is calculated, and the monitored object whose abnormal impact exceeds the preset impact threshold is determined as the target abnormal object. The abnormal impact is calculated based on the position of the monitored object in the propagation sequence, the number of downstream objects affected, and the associated importance of each object.

[0007] In the above embodiment, a directed data dependency graph is established to clarify the interactive relationship of the monitoring object data, and the performance fluctuation sequence arranges the operating performance data in time to provide a time sequence basis for anomaly analysis; the association importance is calculated by weighting direct and indirect associations to quantify the importance of the object in the association network; the propagation consistency factor reflects the degree of propagation of the anomaly in the associated object, and the time closeness factor reflects the time correlation of the associated performance data. The two are weighted with the association importance to obtain the propagation weight, and the constructed propagation sequence clarifies the order of anomaly transmission; combined with the propagation sequence and the dependency graph, the anomaly impact is calculated by monitoring the position of the object in the propagation sequence, the number of downstream objects and the association importance, and the target abnormal object is determined, thereby realizing the tracking of the anomaly propagation process, quantifying the importance of the object and the scope of influence, and improving the accuracy of the platform operation performance evaluation and the accuracy of anomaly positioning.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of collecting the operating performance data of the monitored object and arranging the operating performance data in chronological order to form a performance fluctuation sequence specifically includes: collecting performance indicator data of the monitored object, and the performance indicator data includes at least processor usage data and memory occupancy data; obtaining sampling values ​​of the performance indicator data according to a preset sampling time interval; associating the sampling values ​​with the corresponding sampling time to generate performance indicator time series data, and constructing a performance fluctuation sequence based on the performance indicator time series data, wherein each sequence element includes a time identifier and corresponding performance indicator data.

[0009] In the above embodiment, key performance indicator data such as processor utilization and memory occupancy is collected, sampled at preset sampling intervals, and the sampled values ​​are associated with the sampling time to generate time series data, thereby constructing a performance fluctuation sequence containing time stamps and performance data. Key performance indicators ensure a strong correlation between data and operating status, fixed sampling intervals ensure time series consistency, and the binding of time stamps and performance data makes performance changes traceable. These features work together to accurately capture the dynamic fluctuations of performance over time, providing reliable time series data support for subsequent analysis of anomaly propagation and time correlation, thereby improving the time series accuracy of performance evaluation.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the calculation of the association importance of the monitored object in the directed data dependency graph, the step of weighted calculation of the association importance according to the direct association number and the indirect association number of the monitored object, specifically includes: counting the number of adjacent nodes of the monitored object in the directed data dependency graph, and taking the number of adjacent nodes as the direct association number; traversing the directed data dependency graph, obtaining the number of all reachable nodes of the monitored object, and taking the number of reachable nodes as the indirect association number; multiplying the direct association number by a first preset weight to obtain a direct association score, and multiplying the indirect association number by a second preset weight to obtain an indirect association score; adding the direct association score and the indirect association score to obtain the association importance of the monitored object.

[0011] In the above embodiment, the number of direct connections (neighboring nodes) and indirect connections (reachable nodes) of a monitored object is counted, multiplied by preset weights, and then added together to determine the association importance. The number of direct connections reflects the immediate interactive impact of the object, while the number of indirect connections reflects its radiating influence through other objects. A weighted calculation integrates both. This approach comprehensively considers both the direct and indirect influence of an object in the association network, avoiding the one-sidedness of assessments based solely on direct connections. This provides a precise quantification of importance for subsequent propagation weight calculations and anomaly impact analysis, thereby enhancing the comprehensiveness of object importance assessments.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of calculating the abnormal impact of the monitored object based on the performance fluctuation propagation sequence and the directed data dependency graph, and determining the monitored object whose abnormal impact exceeds the preset impact threshold as the target abnormal object, specifically includes: obtaining the serial number of the monitored object in the performance fluctuation propagation sequence, and determining the position weight value of the monitored object according to the preset position weight calculation rule; determining the number of all downstream nodes of the monitored object in the directed data dependency graph, and multiplying the position weight value by the number of downstream nodes to obtain the initial impact value; calculating the sum of the associated importance of the downstream nodes, and multiplying the initial impact value by the sum of the associated importance to obtain the abnormal impact of the monitored object; judging whether the abnormal impact is greater than the preset impact threshold, and if so, determining the monitored object as the target abnormal object.

[0013] In the above embodiment, a directed data dependency graph is established to clarify the interactive relationship of the monitoring object data, and the performance fluctuation sequence arranges the operating performance data in time to provide a time sequence basis for anomaly analysis; the association importance is calculated by weighting direct and indirect associations to quantify the importance of the object in the association network; the propagation consistency factor reflects the degree of propagation of the anomaly in the associated object, and the time closeness factor reflects the time correlation of the associated performance data. The two are weighted with the association importance to obtain the propagation weight, and the constructed propagation sequence clarifies the order of anomaly transmission; combined with the propagation sequence and the dependency graph, the anomaly impact is calculated by monitoring the position of the object in the propagation sequence, the number of downstream objects and the association importance, and the target abnormal object is determined, thereby realizing the tracking of the anomaly propagation process, quantifying the importance of the object and the scope of influence, and improving the accuracy of the platform operation performance evaluation and the accuracy of anomaly positioning.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after the step of calculating the abnormal impact of the monitored object based on the performance fluctuation propagation sequence and the directed data dependency graph, and determining the monitored object whose abnormal impact exceeds a preset impact threshold as the target abnormal object, the method also includes: obtaining the abnormal point of the performance indicator based on the performance fluctuation sequence of the target abnormal object; in the directed data dependency graph, starting from the target abnormal object, obtaining the performance fluctuation data of the associated objects of the target abnormal object according to the preset hierarchy; performing time alignment on the performance fluctuation data to determine the time relationship of the performance anomaly; determining the abnormal propagation link based on the time relationship and the topological structure of the directed data dependency graph; identifying the abnormal trigger node on the abnormal propagation link, and statistically analyzing the abnormal characteristic parameters of the abnormal trigger node to generate an abnormal analysis report.

[0015] In the above embodiment, key performance indicator data such as processor utilization and memory occupancy is collected, sampled at preset sampling intervals, and the sampled values ​​are associated with the sampling time to generate time series data, thereby constructing a performance fluctuation sequence containing time stamps and performance data. Key performance indicators ensure a strong correlation between data and operating status, fixed sampling intervals ensure time series consistency, and the binding of time stamps and performance data makes performance changes traceable. These features work together to accurately capture the dynamic fluctuations of performance over time, providing reliable time series data support for subsequent analysis of anomaly propagation and time correlation, thereby improving the time series accuracy of performance evaluation.

[0016] In combination with some embodiments of the first aspect, in some embodiments, the step of determining the abnormal propagation link based on the time series relationship and the topological structure of the directed data dependency graph specifically includes: obtaining the fluctuation trend of the performance indicator, and determining an analysis time window based on the fluctuation trend; within the analysis time window, extracting the performance indicator change characteristics of the associated objects of the target abnormal object; constructing the propagation path of the performance fluctuation based on the topological structure of the directed data dependency graph and the performance indicator change characteristics; calculating the propagation order of the associated objects in the propagation path of the performance fluctuation, and determining the abnormal propagation link based on the propagation order.

[0017] In the above example, the analysis time window is determined based on the fluctuation trend of performance indicators, focusing on key periods related to anomalies. Within the window, the performance indicator change characteristics of related objects are extracted to capture anomaly patterns. A propagation path is constructed by combining the dependency graph topology and change characteristics to ensure that the path conforms to the data interaction logic. Finally, the propagation sequence is calculated to determine the anomaly propagation link. These features work together to make the determination of the propagation link more accurate to the actual anomaly spread, improving the accuracy of locating the anomaly propagation path.

[0018] In combination with some embodiments of the first aspect, in some embodiments, the step of identifying the abnormal trigger node on the abnormal propagation link and counting the abnormal characteristic parameters of the abnormal trigger node to generate an abnormal analysis report specifically includes: determining the starting node of the abnormal propagation link as the abnormal trigger node according to the propagation direction of the abnormal propagation link; extracting the performance fluctuation characteristics of the abnormal trigger node, and counting the fluctuation frequency, fluctuation amplitude and fluctuation duration of the abnormal trigger node; using the fluctuation frequency, fluctuation amplitude and fluctuation duration as abnormal characteristic parameters, and generating an abnormal analysis report based on the abnormal characteristic parameters.

[0019] In the above embodiment, the originating node is identified as the anomaly triggering node based on the propagation direction of the anomaly propagation link, thus clarifying the anomaly source. The performance fluctuation characteristics of the node are then extracted, and the frequency, amplitude, and duration of the fluctuations are calculated as anomaly characteristic parameters. The propagation direction ensures accurate positioning of the triggering node, the frequency of fluctuations reflects the frequency of anomalies, the amplitude reflects the severity of the anomaly, and the duration indicates the duration of the impact. These parameters are combined to generate an anomaly analysis report that comprehensively presents the characteristics of the anomaly triggering node, providing a concrete and quantitative basis for anomaly root cause analysis, thereby improving the accuracy of anomaly analysis and the practicality of the report.

[0020] In a second aspect, an embodiment of the present application provides a platform operation performance evaluation system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the platform operation performance evaluation system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions. When the computer program product is run on a platform operation performance evaluation system, the platform operation performance evaluation system is enabled to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium comprising instructions. When the instructions are executed on a platform operation performance evaluation system, the platform operation performance evaluation system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It is understandable that the platform operation performance evaluation system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of the present application. Therefore, the beneficial effects that can be achieved can be referenced to the beneficial effects of the corresponding methods and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. This application clarifies the interactive relationship of monitoring object data by establishing a directed data dependency graph, and arranges the operating performance data in time in the performance fluctuation sequence to provide a time series basis for anomaly analysis; the association importance is calculated by weighting direct and indirect associations to quantify the importance of the object in the association network; the propagation consistency factor reflects the degree of propagation of the anomaly in the associated object, and the time closeness factor reflects the time correlation of the associated performance data. The two are weighted with the association importance to obtain the propagation weight, and the constructed propagation sequence clarifies the order of anomaly transmission; combining the propagation sequence and the dependency graph, the anomaly impact is calculated by monitoring the position of the object in the propagation sequence, the number of downstream objects and the association importance, and the target anomaly object is determined, thereby realizing the tracking of the anomaly propagation process, quantifying the importance of the object and the scope of influence, and improving the accuracy of the platform operation performance evaluation and the accuracy of anomaly positioning.

[0025] 2. This application collects key performance indicator data such as processor usage and memory occupancy, obtains sampling values ​​at preset sampling intervals, associates the sampling values ​​with the sampling time to generate time series data, and constructs a performance fluctuation sequence containing time stamps and performance data. Key performance indicators ensure a strong correlation between data and operating status, fixed sampling intervals ensure time series consistency, and the binding of time stamps and performance data makes performance changes traceable. These features work together to accurately capture the dynamic fluctuation patterns of performance over time, provide reliable time series data support for subsequent analysis of abnormal propagation, time correlation, etc., and improve the time series accuracy of performance evaluation.

[0026] 3. This application calculates the association importance by counting the number of direct connections (adjacent nodes) and indirect connections (reachable nodes) of the monitored object, multiplying each by a preset weight, and then adding the results. The number of direct connections reflects the immediate interactive impact of the object, while the number of indirect connections reflects its radiating effect through other objects. A weighted calculation integrates both. This approach comprehensively considers the direct and indirect influence of an object in the association network, avoiding the one-sidedness of assessments based solely on direct connections. It provides a precise quantification of importance for subsequent propagation weight calculations and abnormal impact analysis, thereby improving the comprehensiveness of object importance assessments. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a platform operation performance evaluation method based on collaborative intelligent analysis in an embodiment of the present application; Figure 2 This is another flowchart of the platform operation performance evaluation method based on collaborative intelligent analysis in an embodiment of the present application; Figure 3 It is a schematic diagram of the structure of a physical device of the platform operation performance evaluation system in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following examples of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application, the singular expressions "a", "an", "above", "the", and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations of one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of the technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, "plurality" means two or more.

[0030] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0031] In a large data center, a cluster consisting of dozens of servers is deployed, responsible for running the company's core business systems. These servers include multiple application servers, database servers, and load balancing servers. Complex data interactions exist between these servers. Application servers need to read data from database servers, load balancing servers need to distribute requests to application servers, and database servers need to synchronize data with backup servers. As business grows, the scale of server clusters continues to expand, and the dependencies between servers become increasingly complex. During peak business hours, some servers often experience response delays. However, operations and maintenance personnel find it difficult to determine whether the server itself is at fault or is affected by other related servers. They are also unable to determine the propagation path and impact range of the anomaly, making troubleshooting difficult and seriously impacting business continuity.

[0032] The data center used traditional performance monitoring methods, setting independent performance thresholds for each server. For example, alarms were triggered when CPU utilization exceeded 85% and memory utilization exceeded 90%. Probes were then used to collect real-time performance data. During a peak business period, an application server triggered a CPU usage alarm. Operations personnel checked and found that the memory usage was normal. They restarted the application server, and response delays were temporarily alleviated. However, shortly thereafter, the application server triggered another alarm, and the database server also experienced slow response times. Operations personnel then checked the database server and discovered that the time spent interacting with the application server had increased significantly. However, because the database server's own CPU utilization did not reach the threshold, no alarm was triggered. Further investigation revealed that the backup server had experienced an anomaly during data synchronization, resulting in resource usage on the database server, which in turn affected the application server. However, existing technologies were unable to correlate the dependencies between servers and only monitored single-point indicators. This made it impossible to track the anomaly's propagation path, leading to a delay in locating the root cause of the problem.

[0033] After implementing this solution, the data center's server performance evaluation improved. The system first obtains data relationships between servers and builds a directed data dependency graph, clearly illustrating the direction of data interaction between servers. It then collects operational performance data, such as CPU usage and memory utilization, from each server and forms a chronological performance fluctuation sequence. By calculating the server's correlation importance, it was found that the load balancing server, directly connected to multiple application servers and indirectly connected to the database server, had a high correlation importance. When analyzing the performance fluctuation sequence, the propagation consistency factor and temporal closeness factor of performance data exceeding the threshold were calculated. It was found that shortly after the application server anomaly, the database server also experienced anomalies, with high propagation consistency and temporal closeness. Based on these factors, propagation weights were calculated to construct a performance fluctuation propagation sequence, clarifying the order in which anomalies propagate from the backup server, through the database server, to the application server. The anomaly impact was then calculated using the directed data dependency graph, identifying the backup server as the target of the anomaly. This enabled operations and maintenance personnel to quickly locate the root cause of the fault and address it promptly, effectively shortening troubleshooting time, preventing further spread of the fault, and ensuring stable business operations.

[0034] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , which is a flow chart of a platform operation performance evaluation method based on collaborative intelligent analysis in an embodiment of the present application.

[0035] S101: Obtain data association relationships between monitoring objects in the platform to be evaluated, and establish a directed data dependency graph based on the data association relationships. Each edge in the directed data dependency graph represents a data interaction relationship between the monitoring objects.

[0036] Among them, the platform to be evaluated refers to the system or architecture collection that needs to be evaluated for operational performance, such as a distributed platform composed of multiple servers and multiple service components; the monitoring object refers to the unit that is monitored in the platform, which may include components with independent functions such as servers, databases, and applications; the data association relationship refers to the dependency relationship formed through data interaction between monitoring objects, such as the call relationship formed by server A sending a data request to server B; the directed data dependency graph is a model that graphically displays the monitoring objects and data association relationships, in which nodes represent monitoring objects and edges represent the direction of data interaction; edges are connecting elements in the directed data dependency graph, used to indicate the direction and existence of data interaction between monitoring objects, for example, an edge from node X to node Y indicates that X sends data to Y.

[0037] This step is typically performed during the initial stage of platform operation, after architectural adjustments, or during regular maintenance. Its purpose is to establish a basic relationship model for subsequent performance evaluation. Specifically, the system first collects data interaction information between all monitored objects through log analysis and interface call logging. For example, server A transmits user data to server B every 5 minutes, and database C receives query requests from application D. This information is then sorted to clarify the direction and content of each monitored object's interaction with other objects, and to determine data relationships. Finally, a directed data dependency graph is constructed based on these relationships. For example, when the monitored objects include application servers, database servers, and cache servers, if the application server needs to write data to the database server, and the database server needs to synchronize data with the cache server, a directed edge "application server → database server → cache server" will be formed in the graph, clearly showing the data flow path.

[0038] In some embodiments, this step can be implemented in a variety of ways: optionally, the system deploys a log collection tool to collect the interaction logs of each monitored object (such as interface call logs, data transmission logs), extract the identifiers and directions of the two interacting parties (such as "object A→object B") from the logs, count all interaction relationships, and then use a graph database (such as Neo4j) to build a directed data dependency graph; optionally, by embedding probes in each monitored object, the communication protocol (such as HTTP requests, TCP connections) and data flow between objects are recorded in real time, and the recorded interaction information is sent to the analysis module. After the analysis module organizes it in the "source object-target object" format, a directed data dependency graph is generated using a visualization tool (such as Gephi). It is understandable that other methods can also be used to achieve the acquisition of data association relationships and the construction of the graph, such as manually combing through the platform architecture document to determine the association relationship and then manually entering it into the system, which is not limited here.

[0039] S102: Collect the operating performance data of the monitored object, and arrange the operating performance data in chronological order to form a performance fluctuation sequence.

[0040] Among them, operation performance data refers to quantitative indicators that reflect the operation status of the monitored object, including at least processor utilization (such as the proportion of CPU core occupancy), memory utilization (such as the ratio of used memory to total memory), and may also include disk IO rate, network bandwidth utilization, etc.; time sequence refers to the order of arrangement according to the time of data generation or collection, such as arranging in sequence of "10:00, 10:01, 10:02"; performance fluctuation sequence refers to an ordered set of operation performance data organized in chronological order, each element contains corresponding time information and performance data, which is used to reflect the performance change trend over time.

[0041] This step is performed continuously or periodically during platform operation, providing a time-series data foundation for subsequent analysis of performance fluctuation patterns and anomaly identification. Specifically, the system first determines the performance metrics to be collected. In addition to processor utilization and memory usage, additional metrics such as database query response time and the number of concurrent application connections can be added based on platform characteristics. The system then sets a sampling interval (e.g., 10 seconds or 1 minute) and periodically collects performance metric values ​​for each monitored object. For example, the CPU utilization of Server A is recorded every 10 seconds. Each collected value is then associated with the corresponding sampling time (e.g., "2025-07-12 10:00:00") to form a "time-performance value" key-value pair. These key-value pairs are then arranged in chronological order to form a performance fluctuation sequence. For example, the CPU utilization sequence for Server A might be "(10:00:00, 60%) → (10:00:10, 65%) → (10:00:20, 80%)," visually demonstrating dynamic performance changes.

[0042] In some embodiments, this step can be implemented in a variety of ways: Optionally, the system uses Prometheus as a monitoring tool, configured to collect indicators such as CPU usage and memory usage of each monitored object at a sampling interval of 30 seconds. Prometheus stores the collected data as time series data according to timestamps. After the system reads this data from Prometheus, it sorts it in ascending order by timestamp to form a performance fluctuation sequence, where each sequence element contains a timestamp (accurate to milliseconds) and the corresponding indicator value. Optionally, by writing a script (such as a Python script), operating system commands (such as the top command or free command in Linux) are used to regularly obtain sampled values ​​of performance indicators. The script binds the sampled values ​​to the current system time (such as obtained through the datetime module) and stores them in a CSV format file (each line contains "time, CPU usage, memory usage"). When the file is subsequently read, it is sorted by the time column to construct a performance fluctuation sequence. It is understandable that other methods can also be used to achieve performance data collection and sequence construction, such as using the Zabbix monitoring system to collect data and then obtain and organize time series data through the API interface. This is not limited here.

[0043] In some embodiments, this step specifically includes: The collection includes performance indicator data of the monitored object, which performance indicator data includes at least processor usage data and memory occupancy data; obtaining sampling values ​​of the performance indicator data according to a preset sampling time interval; associating the sampling values ​​with the corresponding sampling time to generate performance indicator time series data, and constructing a performance fluctuation sequence based on the performance indicator time series data, where each sequence element includes a time identifier and corresponding performance indicator data.

[0044] The monitoring object refers to the monitored functional unit in the platform, such as a server, database, or application. Performance indicator data is quantitative information reflecting the operating status of the monitored object. Processor utilization data refers to the proportion of non-idle time in a unit of time (for example, if a server's processor is in operation for 45 seconds within a minute, the utilization rate is 75%). Memory occupancy data refers to the ratio of used memory capacity to total memory capacity (for example, if the total memory is 16GB and 12GB is used, the utilization rate is 75%). The preset sampling interval is a fixed time period for collecting data (such as 10 seconds or 1 minute) to ensure the temporal regularity of the data. The sampled value is the specific value of the performance indicator collected at each sampling time point. Performance indicator time series data is data that is obtained by binding the sampled value to its corresponding sampling time (for example, "10:00:00, processor utilization 70%, memory utilization 60%"). The performance fluctuation series is a collection of performance indicator time series data arranged in chronological order. Each element contains a time identifier (such as a specific time point) and the corresponding performance indicator data, which is used to reflect the performance trend over time. For example, if you sample an application server at 10-second intervals, the performance fluctuation sequence you get might be [(10:00:00, 70%, 60%), (10:00:10, 75%, 62%), (10:00:20, 80%, 65%)].

[0045] The specific implementation process of this step is as follows: First, determine the monitoring objects (such as the three application servers and two database servers in the cluster) and clearly define the performance indicators to be collected, including at least processor utilization and memory usage. Raw data can be obtained through the server's built-in monitoring tools (such as the Linux system's top command to view processor utilization in real time and the free command to view memory utilization) or the API interface of professional monitoring software (such as Zabbix). Second, set the sampling interval based on the sensitivity of the platform's performance fluctuations (such as 1 second for high-frequency trading platforms and 10 seconds for general business platforms). The system automatically triggers the collection action at this interval, recording the specific values ​​of processor utilization and memory utilization (i.e., sampling values) at each time point. Next, associate each sampling value with the corresponding sampling time (accurate to milliseconds, such as "2025-07-12 14:30:00.000") to generate a performance indicator time series data in the format of (time stamp, processor utilization, memory utilization). Finally, sort all time series data in the order of the time stamps (for example, from early to late) to form a performance fluctuation sequence. For example, the sequence of a database server is [(14:30:00, 65%, 50%), (14:30:10, 68%, 52%), (14:30:20, 72%, 55%)]. Each element fully records the performance status at the corresponding time point, providing structured data for subsequent analysis of performance fluctuations over time.

[0046] S103: Calculate the association importance of the monitoring object in the directed data dependency graph. The association importance is weighted according to the number of direct associations and the number of indirect associations of the monitoring object.

[0047] Among them, association importance refers to a quantitative indicator used to measure the importance of the monitored object in the directed data dependency graph based on the association relationship. The higher the value, the more critical the object is in the data interaction network; the number of direct associations refers to the number of adjacent nodes directly connected to the monitored object in the directed data dependency graph. For example, if the monitored object A has direct data interaction with objects B and C, then the number of direct associations of A is 2; the number of indirect associations refers to the number of all nodes that the monitored object can indirectly reach through other nodes in the directed data dependency graph. For example, if object A is directly associated with B, and B is associated with C, then the number of indirect associations of A at least includes C; weighted calculation refers to a calculation method that assigns preset weights to the number of direct associations and the number of indirect associations respectively, and then obtains a comprehensive result through multiplication and addition operations. For example, a weight of 0.6 is assigned to the number of direct associations and a weight of 0.4 is assigned to the number of indirect associations, and then calculation is performed.

[0048] This step is performed after the directed data dependency graph is constructed. Its purpose is to quantify the importance of each monitored object in the association network and provide a basis for the subsequent calculation of propagation weight and anomaly impact. Specifically, the system first traverses the directed data dependency graph, counts the number of adjacent nodes of each monitored object, and determines it as the number of direct associations. For example, in a graph containing "application server → database server → cache server", the number of direct associations of the database server is 2 (application server and cache server); then, the graph is traversed using a depth-first or breadth-first algorithm to obtain the number of all reachable nodes of the monitored object as the number of indirect associations. For example, the above-mentioned database server can be associated with the log server through the cache server, so its indirect association number includes the log server; then, according to the preset first and second weights (such as the first weight is 0.7 and the second weight is 0.3), the number of direct associations and the number of indirect associations are weighted respectively to obtain the direct association score and the indirect association score; finally, the two scores are added together to obtain the association importance of the monitored object. For example, if the number of direct associations is 3, the first weight is 0.6, and the number of indirect associations is 5, and the second weight is 0.4, then the association importance = 3×0.6+5×0.4=1.8+2=3.8.

[0049] In some embodiments, this step can be implemented in a variety of ways: optionally, the system uses a graph traversal tool (such as NetworkX) to load a directed data dependency graph, calls the tool's "neighbors" method to obtain the adjacent nodes of each monitored object, and counts the number to obtain the number of direct associations; then uses the "reachable_nodes" method to obtain all reachable nodes, and counts the number to obtain the number of indirect associations; then reads the first preset weight (such as 0.5) and the second preset weight (such as 0.5) from the configuration file, multiplies them by the direct and indirect association numbers respectively, and then sums them to obtain the association importance; optionally, the graph traversal is implemented by programming, for example, using Python code to traverse the adjacency list of the directed graph, and for each monitored object, cyclically counts the number of directly connected nodes as the number of direct associations; then recursively traverses all reachable nodes, records and deduplicates, and counts the number as the number of indirect associations; finally, it is calculated by the formula "association importance = number of direct associations × w1 + number of indirect associations × w2" (where w1 and w2 are preset weights, such as w1=0.6, w2=0.4). It is understandable that other methods may be used to calculate the association importance, such as adjusting the weight value in combination with the interaction frequency of the nodes, which is not limited here.

[0050] In some embodiments, this step specifically includes: counting the number of adjacent nodes of the monitored object in the directed data dependency graph, and taking the number of adjacent nodes as the number of direct associations; traversing the directed data dependency graph, obtaining the number of all reachable nodes of the monitored object, and taking the number of reachable nodes as the number of indirect associations; multiplying the number of direct associations by the first preset weight to obtain a direct association score, and multiplying the number of indirect associations by the second preset weight to obtain an indirect association score; adding the direct association score and the indirect association score to obtain the association importance of the monitored object.

[0051] A directed data dependency graph refers to a graphical model in which nodes represent monitored objects and directed edges represent the direction of data interaction between objects. For example, the edge "Server A → Server B" indicates that A sends data to B. Adjacent nodes refer to nodes directly connected to the current monitored object in the directed data dependency graph (including upstream nodes that send data and downstream nodes that receive data). For example, if server A is directly connected to B and C, then B and C are adjacent nodes of A. The number of direct associations refers to the total number of adjacent nodes. Reachable nodes refer to all nodes (excluding the current monitored object) that can be indirectly reached from the current monitored object via directed edges. For example, in the case of A → B → C, C is a reachable node of A. The number of indirect associations refers to the total number of reachable nodes. The first preset weight and the second preset weight are pre-set coefficients used for weighted calculation (the sum of the two is usually 1, such as the first weight is 0.6 and the second weight is 0.4). The direct association score is the product of the number of direct associations and the first preset weight. The indirect association score is the product of the number of indirect associations and the second preset weight. The association importance is the sum of the direct association score and the indirect association score, which is used to quantify the importance of the monitored object in the dependency graph. For example, the number of direct associations of monitored object A is 2 (adjacent nodes B and C), the number of indirect associations is 3 (reachable nodes D, E, and F), the first weight is 0.5, and the second weight is 0.5. Then the direct association score = 2 × 0.5 = 1, the indirect association score = 3 × 0.5 = 1.5, and the association importance = 1 + 1.5 = 2.5.

[0052] The specific implementation process of this step is as follows: For each monitored object (node) in the directed data dependency graph, all edges in the graph are traversed and the number of adjacent nodes directly connected to the node (including upstream nodes pointed to by incoming edges and downstream nodes pointed to by outgoing edges) is counted. This number is determined as the number of direct connections. For example, if node X has three incoming edges (from A, B, and C) and two outgoing edges (to D and E), it has five adjacent nodes, and the number of direct connections is five. The directed data dependency graph is traversed using a depth-first search (DFS) or breadth-first search (BFS): Starting from the current monitored object, all reachable nodes are visited along the directed edges in sequence, recording and removing duplicates (to avoid double counting). The total number of these nodes is counted, which is the number of indirect connections. For example, node Y can reach F, G, and H via the path Y→F→G→H, and can reach I via Y→I. Therefore, there are four reachable nodes, and the number of indirect connections is four. The system reads the first preset weight (e.g., 0.6) and the second preset weight (e.g., 0.4) from the system configuration. The number of direct connections is multiplied by the first preset weight to obtain the direct connection score (e.g., if the number of direct connections is 3, then 3 × 0.6 = 1.8). The number of indirect connections is multiplied by the second preset weight to obtain the indirect connection score (e.g., if the number of indirect connections is 5, then 5 × 0.4 = 2). The direct and indirect connection scores are added together to obtain the association importance of the monitored object (e.g., 1.8 + 2 = 3.8). This approach considers both the direct influence of the monitored object (direct connection) and the radiation effect generated through indirect relationships (indirect connection), achieving a comprehensive quantification of node importance.

[0053] S104. Analyze the performance fluctuation sequence, calculate the propagation consistency factor of the performance data exceeding the preset performance threshold and the time closeness factor of the associated performance data in the performance fluctuation sequence, wherein the propagation consistency factor indicates the degree of propagation of the performance data between associated objects, and the time closeness factor indicates the degree of time correlation of the associated performance data.

[0054] Among them, the propagation consistency factor refers to a quantitative indicator used to represent the degree of propagation of performance data that exceeds the preset performance threshold among associated objects. The higher the value, the stronger the consistency of the propagation of abnormal data in the associated objects. For example, if an anomaly occurs in 80% of the associated objects, the propagation consistency factor may be 0.8; the time closeness factor refers to a quantitative indicator used to represent the degree of temporal correlation of associated performance data. The higher the value, the closer the performance data anomalies of the associated objects are in time. For example, if the time difference between the occurrence of anomalies of two associated objects is 1 second, the time closeness factor may be higher than the case where the time difference is 10 seconds; the preset performance threshold refers to the critical value set in advance to determine whether the performance data is abnormal. For example, the preset threshold of CPU utilization is 85%, and exceeding it is considered an anomaly; associated performance data refers to the performance data of monitoring objects that have data association relationships in a directed data dependency graph. For example, if objects A and B have data interaction, the performance data of A and B are mutually associated performance data.

[0055] This step is performed after the performance fluctuation sequence is formed. Its purpose is to quantify the propagation characteristics and temporal correlation characteristics of the abnormal data, providing a basis for the subsequent construction of the performance fluctuation propagation sequence. Specifically, the system first analyzes the performance fluctuation sequence element by element, filtering out all performance data that exceeds the preset performance threshold (i.e., abnormal data). Then, for these abnormal data, the proportion of their occurrence in the associated objects is counted (for example, what proportion of the associated objects also have similar anomalies). This is used to calculate the propagation consistency factor, with higher proportions indicating larger factor values. Simultaneously, the time points at which the anomalies occurred in the associated performance data are extracted, and the time difference is calculated (for example, the difference between the abnormal time of object A and the abnormal time of object B). The time closeness factor is obtained by taking the inverse of the time difference or normalizing it. The smaller the time difference, the larger the factor value, reflecting the degree of temporal correlation.

[0056] In some embodiments, this step can be implemented in a variety of ways: Optionally, the system extracts abnormal data exceeding a preset threshold from the performance fluctuation sequence, determines all associated objects of the monitoring object to which the data belongs, counts the proportion of the number of associated objects with similar abnormalities (e.g., all exceeding the CPU usage threshold), and uses the proportion as a propagation consistency factor (e.g., there are 5 associated objects, 4 of which have abnormalities, then the factor is 4 / 5=0.8); at the same time, records the abnormality occurrence time of the abnormal object and the associated objects, calculates the time difference Δt, and uses the formula "time closeness factor = 1 / ( 1+Δt)" (e.g., if Δt=2 seconds, then the factor = 1 / 3≈0.33). The smaller Δt is, the closer the factor is to 1. Optionally, the propagation consistency factor is obtained by calculating the distribution similarity of abnormal data in associated objects (e.g., using a cosine similarity algorithm). The higher the similarity, the stronger the propagation consistency. For the time closeness factor, a fixed time window (e.g., 30 seconds) is set, and the overlap duration of the abnormal time of the associated performance data within the window is counted. The ratio of the overlap duration to the total window duration is used as the factor (e.g., if the overlap duration is 20 seconds, then the factor = 20 / 30≈0.67). It is understandable that other methods can also be used to calculate the two factors, such as obtaining the consistency factor by fitting the propagation law through a machine learning model, which is not limited here.

[0057] S105. Calculate the propagation weight value between any two monitored objects and construct a performance fluctuation propagation sequence of the monitored objects based on the propagation weight value. The propagation weight value is obtained by weighted calculation of the associated importance, the propagation consistency factor and the time closeness factor. The propagation sequence represents the transmission order of performance anomalies between each monitored object.

[0058] Among them, the propagation weight value refers to a quantitative indicator used to measure the possibility of performance anomaly transmission between any two monitored objects. It is obtained by weighted calculation of the association importance, propagation consistency factor and temporal closeness factor. The higher the value, the greater the possibility of the anomaly transmitting from one object to another. The performance fluctuation propagation sequence refers to the sequence of monitoring objects arranged from high to low according to the possibility of anomaly transmission, which is used to represent the transmission order of performance anomalies between each monitored object. For example, the sequence "object A → object B → object C" means that the anomaly is more likely to occur first in A, then transmit to B, and finally to C. Weighted calculation refers to the process of assigning preset weights (such as 0.4, 0.3, and 0.3) to the association importance, propagation consistency factor and temporal closeness factor respectively, and obtaining the comprehensive result through multiplication and addition operations.

[0059] This step is performed after calculating the correlation importance, propagation consistency factor, and temporal closeness factor. Its purpose is to prioritize anomaly propagation by integrating multiple metrics, providing a basis for subsequently locating key anomaly objects. Specifically, the system first extracts the correlation importance (e.g., the correlation importance of X to Y is 3.8), propagation consistency factor (e.g., 0.8), and temporal closeness factor (e.g., 0.9) of any two monitored objects with data associations (e.g., object X and object Y). Then, based on preset weights (e.g., correlation importance weight 0.4, propagation consistency factor 0.3, temporal closeness factor 0.3), the propagation weight between the two objects is calculated using the formula: "Propagation weight = correlation importance × 0.4 + propagation consistency factor × 0.3 + temporal closeness factor × 0.3." Finally, the propagation weights between all monitored objects are sorted, and the order of anomaly propagation is determined from high to low weight, forming a performance fluctuation propagation sequence. For example, if the propagation weight between objects A and B is 0.8, and between B and C is 0.6, the sequence might be "A→B→C."

[0060] In some embodiments, this step can be implemented in various ways: Optionally, the system reads preset weights (w1=0.5, w2=0.3, and w3=0.2, corresponding to association importance, propagation consistency factor, and temporal closeness factor, respectively) from a configuration file. For any two associated monitoring objects, the system then uses the three metrics calculated in the previous steps and substitutes them into the formula: propagation weight = w1 × association importance + w2 × propagation consistency factor + w3 × temporal closeness factor. The propagation weights between all objects are then stored in a matrix and processed using a topological sorting algorithm (such as the Kahn algorithm) to generate a performance fluctuation propagation sequence from highest to lowest weight. Optionally, the system first assigns differentiated weights to different types of monitoring objects (e.g., servers and databases) (e.g., servers have a higher association importance weight), then calculates propagation weights for each pair of objects. Then, based on the edge direction of the directed data dependency graph, the system retains only those propagation weights that align with the data flow direction (e.g., only calculating weights from upstream objects to downstream objects). These weights are sorted in descending order to construct a performance fluctuation propagation sequence, ensuring that the sequence aligns with the actual direction of data interaction. It is understandable that this step can also be implemented in other ways, such as fitting historical data through a machine learning model, dynamically adjusting the weights of each factor, and then calculating the propagation weight value, which is not limited here.

[0061] S106. Calculate the abnormal impact of the monitored object based on the performance fluctuation propagation sequence and the directed data dependency graph, and determine the monitored object whose abnormal impact exceeds the preset impact threshold as the target abnormal object. The abnormal impact is calculated based on the position of the monitored object in the propagation sequence, the number of downstream objects affected, and the associated importance of each object.

[0062] Among them, the abnormal impact refers to a quantitative indicator used to measure the impact of the performance abnormality of the monitored object on the operation of the entire platform. It comprehensively considers the position of the monitored object in the propagation sequence, the number of downstream objects affected, and the associated importance of each object. The higher the value, the more significant the impact; the target abnormal object refers to the monitored object whose abnormal impact exceeds the preset impact threshold, that is, the key abnormal source that has a greater impact on the platform operation; the preset impact threshold refers to the critical value set in advance to determine whether the abnormal impact is significant, such as setting the threshold to 80 (full score 100) based on the platform's historical fault data; the downstream object refers to the object directly or indirectly pointed to by the current monitored object in the directed data dependency graph, that is, the associated object that receives the current object data, for example, in "object A→object B→object C", B and C are downstream objects of A.

[0063] This step is performed after obtaining the performance fluctuation propagation sequence and directed data dependency graph. Its purpose is to filter out the key objects with the greatest impact from multiple abnormal objects, providing a precise target for troubleshooting. Specifically, the system first determines the position of each monitored object in the performance fluctuation propagation sequence (for example, the object at the first position in the sequence will experience an anomaly earlier than the object at the fifth position), sets a weight based on the position (for example, the weight for the first position is 1.0, the weight for the second position is 0.8, and so on); then, in the directed data dependency graph, the number of downstream objects of the object (including direct and indirect downstream objects) is counted; the sum of the associated importance of all downstream objects is calculated; finally, the anomaly impact is calculated using the formula (for example, "anomaly impact = position weight × number of downstream objects × sum of downstream associated importance") and compared with the preset threshold. Objects exceeding the threshold are identified as target anomaly objects.

[0064] In some embodiments, this step can be implemented in a variety of ways: optionally, the system assigns linear weights to positions in the propagation sequence (such as the weight of the nth position is 1-(n-1) / N, where N is the total length of the sequence), counts the number of downstream objects of the monitored object (such as traversing all downstream nodes through a depth-first search of the graph), and calculates the sum of the associated importance of the downstream objects; then calculates using the formula "abnormal impact = position weight × number of downstream objects × sum of downstream associated importance", for example, if the position weight is 0.9, the number of downstream objects is 5, and the sum of associated importance is 20, then the abnormal impact = 0.9 × 5 × 20 = 90. If the preset threshold is 70, the object is determined to be a target abnormal object. Optionally, the system uses the Analytic Hierarchy Process (AHP) to determine weights for location, downstream quantity, and relevance importance (e.g., a weight of 0.4 for location, 0.3 for downstream quantity, and 0.3 for relevance importance). For each monitored object, the system calculates standardized values ​​for these three indicators (e.g., normalizing location to a value between 0 and 1). The system then calculates the impact of anomaly using the formula "Anomaly Impact = 0.4 × Normalized Location Value + 0.3 × Normalized Downstream Quantity Value + 0.3 × Normalized Downstream Relevance Importance Value." The result is compared with a preset threshold (e.g., 0.7). Objects exceeding the threshold are designated as anomalous. It is understood that other methods can also be used to implement this step, such as adjusting the calculation method based on platform business priorities (e.g., payment-related objects have a higher impact weight), which is not limited here.

[0065] In some embodiments, this step specifically includes: Obtain the serial number of the monitored object in the performance fluctuation propagation sequence, and determine the position weight value of the monitored object according to the preset position weight calculation rule; determine the number of all downstream nodes of the monitored object in the directed data dependency graph, and multiply the position weight value by the number of downstream nodes to obtain the initial impact value; calculate the sum of the associated importance of the downstream nodes, multiply the initial impact value by the sum of the associated importance to obtain the abnormal impact of the monitored object; determine whether the abnormal impact is greater than the preset impact threshold. If so, determine the monitored object as a target abnormal object.

[0066] Among them, the performance fluctuation propagation sequence refers to the sequence that represents the order in which performance anomalies are transmitted between each monitored object, for example, [object A, object B, object C] means that the anomaly first appears in A, then transmits to B, and finally to C; the sequence number refers to the position number of the monitored object in the performance fluctuation propagation sequence, such as the sequence number of object A in the above sequence is 1, and the sequence number of object B is 2; the preset position weight calculation rule refers to a pre-set rule that determines the position weight value based on the sequence number. Generally, the earlier the sequence number (the earlier the anomaly appears), the higher the weight value, for example, the rule is "position weight value = 1-(sequence number-1) / total length of the sequence"; the position weight value refers to the value calculated according to the above rules that reflects the importance of the position of the monitored object in the propagation sequence, and the range is usually between 0 and 1; the directed data dependency graph refers to a graph in which nodes represent monitored objects and directed edges represent the direction of data interaction. Model; downstream nodes refer to nodes directly or indirectly pointed to by the current monitored object in the directed data dependency graph (i.e., nodes that receive the current object data), such as "object M→object N→object P", where N and P are downstream nodes of M; the number of downstream nodes refers to the total number of all downstream nodes; the initial impact value refers to the product of the position weight value and the number of downstream nodes, reflecting the combined influence of the position and the number of downstream nodes; the sum of the associated importances refers to the sum of the associated importances of all downstream nodes (the associated importance is the indicator calculated in S103); the abnormal impact refers to a quantitative indicator reflecting the degree of impact of the abnormality of the monitored object on the platform, which is obtained by multiplying the initial impact value and the sum of the associated importances; the preset impact threshold refers to a pre-set critical value for judging whether the abnormal impact is significant, such as 80; the target abnormal object refers to a monitored object whose abnormal impact is greater than the preset impact threshold.

[0067] For example, the performance fluctuation propagation sequence is [X, Y, Z] (total length 3), and X's sequence number is 1. According to the rule "position weight = 1-(sequence number-1) / 3", its position weight = 1-0 / 3 = 1. In the directed data dependency graph, X's downstream nodes are Y, Z, and W, a total of 3 (number of downstream nodes = 3). Y's relevance importance is 2.5, Z's is 3.0, and W's is 1.5. The sum of the relevance importances of the downstream nodes is 2.5 + 3.0 + 1.5 = 7. Therefore, the initial impact value is 1 × 3 = 3, and the abnormal impact is 3 × 7 = 21. If the preset impact threshold is 15, X's abnormal impact 21> 15, so X is determined to be the target abnormal object.

[0068] The specific implementation process of this step is as follows: Obtain the position number (i.e., sequence number) of the monitored object in the performance fluctuation propagation sequence and substitute it into the preset position weight calculation rule (e.g., "position weight = 1 - (sequence number - 1) / total sequence length") to calculate the position weight. In the directed data dependency graph, traverse all directed edges from the current monitored object to determine all nodes that directly and indirectly receive its data (i.e., downstream nodes). Count these nodes to obtain the total number of downstream nodes. Multiply the position weight by the number of downstream nodes to obtain the initial impact value. Extract the association importance of all downstream nodes (calculated in S103) and add these values ​​to obtain the sum of the association importance of the downstream nodes. Multiply the initial impact value by the sum of the association importances to obtain the anomaly impact of the monitored object. Compare the anomaly impact with a preset impact threshold. If the anomaly impact exceeds the threshold, the monitored object is identified as a target anomaly. This approach comprehensively considers the order in which the anomaly occurs, the scope of impact, and the importance of the affected object to accurately locate the source of anomalies that significantly impact platform operations.

[0069] The following is a more detailed description of the process of the method provided by this implementation. Figure 2 , which is another flow chart of the platform operation performance evaluation method based on collaborative intelligent analysis in an embodiment of the present application.

[0070] S201. Obtain abnormal points of performance indicators based on the performance fluctuation sequence of the target abnormal object.

[0071] The target anomaly object refers to a monitored object whose anomaly impact exceeds the preset impact threshold, i.e., an anomaly source with a significant impact on platform operations. The performance fluctuation sequence is an ordered set of the monitored object's operational performance data, arranged in chronological order. Each element contains a timestamp and corresponding performance indicator data. Performance indicators are parameters that reflect the monitored object's operational status, including at least processor utilization and memory usage. Anomaly points are time points and corresponding performance data at which the value of a performance indicator exceeds the preset threshold. For example, if the target anomaly object is a database server, and its performance fluctuation sequence contains elements such as (10:05:00, memory utilization 92%) and (10:05:10, memory utilization 93%), and the preset memory utilization threshold is 90%, then these two time points and their corresponding memory utilization data are considered anomaly points.

[0072] The specific implementation process of this step is as follows: First, identify the performance fluctuation sequence corresponding to the target abnormal object. This sequence contains the performance indicator data of this object at different times. Second, set a preset threshold for each performance indicator, such as 85% for processor utilization and 90% for memory usage. Then, traverse each element in the performance fluctuation sequence, extract the performance indicator data within the element, and compare it with the corresponding preset threshold. Finally, filter out all elements whose performance indicator data exceeds the preset threshold. These elements are considered performance indicator outliers. For example, if the target abnormal object is a cache server, and its performance fluctuation sequence is (9:50:00, CPU utilization 80%), (9:50:10, CPU utilization 87%), (9:50:20, CPU utilization 89%), and the preset CPU utilization threshold is 85%, then after traversing, (9:50:10, 87%) and (9:50:20, 89%) are identified as outliers.

[0073] S202 . In the directed data dependency graph, starting from the target abnormal object, obtain performance fluctuation data of the associated objects of the target abnormal object according to a preset hierarchy.

[0074] A directed data dependency graph refers to a graphical model in which nodes represent monitored objects and directed edges represent the direction of data interaction between monitored objects. The preset level refers to the pre-set number of levels for obtaining associated objects, which limits the scope of associated objects. For example, one level indicates obtaining only directly associated objects, while two levels indicate obtaining directly associated objects and their associated objects. Associated objects refer to monitored objects in the directed data dependency graph that have direct or indirect data interaction relationships with the target anomaly object, including upstream objects (sending data to the target object) and downstream objects (receiving data sent by the target object). Performance fluctuation data refers to the performance fluctuation sequence formed by chronologically arranging the operating performance data of associated objects. For example, if the target anomaly object is an application server, and its first-level associated objects in the directed data dependency graph are the load balancing server (upstream) and the database server (downstream), and the preset level is 1, the performance fluctuation sequence of the load balancing server and the database server will be obtained as performance fluctuation data.

[0075] The specific implementation process of this step is as follows: first, locate the node corresponding to the target abnormal object in the directed data dependency graph; second, based on the preset hierarchy (e.g., 2 layers), starting from the target abnormal object node, search for the first-layer associated objects, i.e., the upstream and downstream nodes directly connected to the target abnormal object; then, starting from the first-layer associated objects, search for the second-layer associated objects, i.e., the direct upstream and downstream nodes of the first-layer associated objects; then, collect all associated objects that meet the preset hierarchy; finally, obtain the performance fluctuation sequence of these associated objects, which is the required performance fluctuation data. For example, if the target abnormal object is A, and the first-layer associated objects of A in the directed data dependency graph are B (A→B) and C (C→A), and the first-layer associated object of B is D (B→D), and the preset hierarchy is 2, then the performance fluctuation sequence of B, C (1st layer), and D (2nd layer) needs to be obtained.

[0076] S203: Perform time series alignment on the performance fluctuation data to determine the time series relationship of the performance anomaly.

[0077] Temporal alignment involves adjusting the performance fluctuation data of different associated objects to a unified time base, making the performance data of each object directly comparable across time. The temporal relationship of performance anomalies refers to the temporal order of the anomalies of associated objects, specifically which object experienced the anomaly first and which object experienced the anomaly later. For example, if the anomaly of associated object B occurred at 14:30:02 and the anomaly of associated object C occurred at 14:30:00, temporal alignment can confirm that the anomaly of object C occurred before that of object B, and the temporal relationship between the two is C → B.

[0078] The specific implementation process of this step is as follows: First, collect performance fluctuation data for the target anomaly object and its associated objects, and extract the time identifier (such as a timestamp) from each data element. Second, convert all timestamps to a unified format (such as a Unix timestamp, accurate to milliseconds) to ensure consistent time bases. Then, sort all performance fluctuation data based on the timestamps, so that the performance data of different objects at the same time point form a corresponding relationship. Next, filter all anomalies from the aligned performance fluctuation data and extract their timestamps. Finally, compare the timestamps of these anomalies. The smaller the timestamp value, the earlier the anomaly occurred. This determines the order of the anomalies for each object, i.e., the temporal relationship of the performance anomalies. For example, if the timestamp of the anomaly for object D is 1689000000000 (15:00:00), and the timestamp of the anomaly for object E is 1689000001000 (15:00:01), after temporal alignment, it can be determined that the anomaly for object D occurs before that for object E, resulting in a temporal relationship of D → E.

[0079] S204: Obtain the fluctuation trend of the performance indicator, and determine an analysis time window according to the fluctuation trend.

[0080] The fluctuation trend of a performance indicator (such as CPU utilization and memory usage) refers to the overall direction of change over a period of time, including patterns such as rising, falling, stable, and sudden rises and falls. The analysis time window is a specific time interval defined for focused analysis of the propagation of performance anomalies, covering the critical stages of anomaly occurrence, development, and spread. For example, if a server's CPU utilization suddenly rises from 60% to 95% between 10:00 and 10:05 and continues to fluctuate, then gradually declines after 10:10, its fluctuation trend is "sudden rise-fluctuation-decline." The corresponding analysis time window can be set to 10:00-10:10.

[0081] The specific implementation process of this step is as follows: First, based on the performance fluctuation sequence, by drawing a time series curve or calculating the sliding window mean, the fluctuation trend of the performance indicator is identified, such as the start time of the anomaly (the first time the indicator exceeds the threshold), the peak time (the highest value of the indicator), and the plateau time (the time when the indicator stabilizes). Then, starting from the anomaly start time, a certain buffer time is reserved forward (such as 30 seconds to capture the precursor to the anomaly), and ending at the plateau time, a certain buffer time is reserved backward (such as 1 minute to capture the aftermath of the anomaly). The interval between the start buffer time and the end buffer time is determined as the analysis time window. For example, if the anomaly starts at 10:00:00, peaks at 10:02:30, and plateaus at 10:05:00, the analysis time window can be set to 09:59:30-10:06:00.

[0082] S205: Extracting performance indicator change characteristics of objects associated with the target abnormal object within the analysis time window.

[0083] The analysis time window refers to the time interval determined in S204 for centralized analysis. The associated objects of the target anomaly object refer to the monitored objects in the directed data dependency graph that have direct or indirect data interactions with the target anomaly object (e.g., upstream objects providing data, downstream objects receiving data). The performance indicator change characteristics refer to the specific change patterns of the associated object's performance indicators (e.g., CPU usage, response time) within the analysis time window, including the magnitude of the change, the rate of change, and the mutation point. For example, if the CPU usage of a certain associated object increases from 70% to 90% within 10 seconds within the analysis time window, this "substantial increase in a short period of time" is its change characteristic.

[0084] The specific implementation process of this step is as follows: First, locate the analysis time window and extract the performance fluctuation sequence of all associated objects within that window. Then, analyze the performance indicator data of each associated object segment by segment and calculate characteristic parameters: change amplitude (the difference between the maximum and minimum values ​​of the indicator within the window), change rate (the amount of change in the indicator per unit time, such as a 5% increase per second), mutation point (the time point when the indicator change rate exceeds the preset threshold, such as a 10% increase in 1 second), and duration of the abnormality (the cumulative time the indicator exceeds the threshold). Finally, these parameters are integrated to form the performance indicator change characteristic of each associated object. For example, the CPU usage of associated object D increased from 65% to 92% within the window (a 27% amplitude), with an average increase of 3% per second (rate), a sudden increase at 10:01:05 (mutation point), and the duration of the abnormality exceeding the threshold for 4 minutes (persistent abnormality). These parameters together constitute its change characteristic.

[0085] S206. Construct a propagation path of performance fluctuations based on the topological structure of the directed data dependency graph and the characteristics of performance indicator changes.

[0086] The topology of a directed data dependency graph refers to the connection between nodes (monitored objects) and the data flow relationship within the graph, including adjacency (direct connections) and hierarchical relationships (upstream and downstream) between nodes. Performance indicator change characteristics refer to the performance change patterns of the associated objects extracted in S205 within the analysis time window. The propagation path of performance fluctuations refers to the specific route along which performance anomalies propagate from the initial object to other associated objects, reflecting the order and direction of anomaly diffusion. For example, in the topology "X→Y→Z," if X first exhibits a "sudden rise" feature, followed by similar features in Y and Z, the propagation path is X→Y→Z.

[0087] The specific implementation process of this step is as follows: First, based on the topological structure of the directed data dependency graph, the potential propagation direction between related objects is determined (for example, only along directed edges from upstream to downstream, excluding reverse paths). Then, the time series of the performance indicator change characteristics of each related object are compared, and the object with the earliest feature appearance time is identified as the potential source, and the object with the later feature appearance time is identified as the subsequent node. Next, the potential path is verified to see if it conforms to the topological structure (for example, whether there is a directed edge connecting the source to the subsequent node). If so, it is included in the propagation path. Finally, all qualified nodes are connected in series according to the order of feature appearance time to form the propagation path of the performance fluctuation. For example, if the topological structure is A→B→C, and the characteristics show that the anomaly of A occurred at 10:00, B at 10:00:10, and C at 10:00:20, and all have directed edges, then the propagation path is A→B→C.

[0088] S207: Calculate the propagation order of the associated objects in the propagation path of the performance fluctuation, and determine the abnormal propagation link according to the propagation order.

[0089] The propagation order refers to the chronological order in which anomalies occur in associated objects along the performance fluctuation propagation path, clarifying the process by which anomalies are transmitted between objects. The anomaly propagation link is a chain of propagation orders, reflecting the complete path of anomalies from source to destination. It is a concrete representation of the performance fluctuation propagation path. For example, if the anomaly of associated object D occurs earlier than that of object E, and E occurs earlier than that of object F, the propagation order is D→E→F, and the corresponding anomaly propagation link is D→E→F.

[0090] The specific implementation process of this step is as follows: First, extract the anomaly timestamps of all associated objects from the performance fluctuation propagation path (i.e., the time when the performance metric first exceeded the threshold). Then, sort these timestamps from smallest to largest to determine the order of anomalies among the associated objects, i.e., the propagation order. Finally, according to the propagation order, connect the associated objects with arrows to form a chain, which is the anomaly propagation link. For example, if the propagation path contains objects M, N, and P, with anomaly timestamps of 15:01:03, 15:01:05, and 15:01:01, respectively, the propagation order after sorting is P→M→N, so the anomaly propagation link is determined to be P→M→N.

[0091] S208: Determine the starting node of the abnormal propagation link as the abnormality triggering node according to the propagation direction of the abnormal propagation link.

[0092] The exception propagation link refers to the chain of associated objects connected in a propagation order, as determined in S207. The propagation direction refers to the direction in which the exception propagates from the object that first experiences the exception to the object that subsequently experiences the exception, i.e., along the propagation order. The starting node refers to the object that first experiences the exception in the exception propagation link and is the starting point of the exception propagation. The exception triggering node refers to the source object that triggers the entire exception propagation process, i.e., the starting node. For example, if the exception propagation link is A→B→C, the propagation direction is A to B and then to C, and the starting node is A, then A is the exception triggering node.

[0093] The specific implementation process of this step is as follows: first, the components of the anomaly propagation chain and the order of the objects are determined; then, based on the propagation order, the chain's propagation direction is determined (e.g., from left to right); then, the object at the front (first position) of the chain is located, i.e., the starting node; finally, this starting node is identified as the anomaly triggering node, as it is the initial initiator of the entire anomaly propagation process. For example, if the anomaly propagation chain is database server → application server → load balancing server, and the propagation direction is from the database server to the load balancing server, the starting node is the database server, so the database server is identified as the anomaly triggering node.

[0094] S209: extracting the performance fluctuation characteristics of the abnormal triggering node, and counting the fluctuation frequency, fluctuation amplitude and fluctuation duration of the abnormal triggering node.

[0095] Among them, the abnormal trigger node refers to the source object that caused the abnormal propagation determined in S208; the performance fluctuation characteristics refer to the overall change pattern of the performance indicators (such as CPU usage and memory occupancy) of the abnormal trigger node within the analysis time window, including the frequency, amplitude, and duration of the fluctuation; the fluctuation frequency refers to the number of times the performance indicator exceeds the preset threshold within the analysis time window; the fluctuation amplitude refers to the maximum difference between the performance indicator and the threshold, that is, the difference between the maximum value of the indicator and the threshold; the fluctuation duration refers to the total length of time that the performance indicator continues to exceed the threshold within the analysis time window. For example, if the abnormal trigger node is a server, its CPU usage exceeds the 85% threshold three times within 10 minutes, reaching a maximum of 95%, and the cumulative exceeding time is 5 minutes, then the fluctuation frequency is 3 times, the fluctuation amplitude is 10%, and the fluctuation duration is 5 minutes.

[0096] The specific implementation process of this step is as follows: First, extract the performance fluctuation sequence of the anomaly triggering node within the analysis time window; then count the number of times the performance metric exceeds the preset threshold in this sequence to obtain the fluctuation frequency; calculate the difference between the metric's maximum value and the preset threshold (for example, if the threshold is 85% and the maximum value is 95%, the amplitude is 10%) to obtain the fluctuation amplitude; accumulate all time periods exceeding the threshold (for example, 10:00-10:02 and 10:03-10:06, a total of 3 + 3 = 6 minutes) to obtain the fluctuation duration; finally, combine these parameters to form the performance fluctuation characteristics of the anomaly triggering node. For example, if the performance fluctuation sequence of the anomaly triggering node shows that the CPU usage exceeds the threshold four times within 30 minutes, the highest value differs from the threshold by 15%, and the cumulative duration of the exceedance is 12 minutes, the statistical results are: fluctuation frequency 4 times, fluctuation amplitude 15%, and fluctuation duration 12 minutes.

[0097] S210: Use the fluctuation frequency, fluctuation amplitude, and fluctuation duration as abnormal characteristic parameters, and generate an abnormal analysis report based on the abnormal characteristic parameters.

[0098] The fluctuation frequency refers to the number of times the performance indicator of the anomaly triggering node exceeds the preset threshold within the analysis time window, which is used to reflect the frequency of anomalies. The fluctuation amplitude refers to the difference between the maximum value of the anomaly triggering node's performance indicator and the preset threshold, which is used to reflect the severity of the anomaly. The fluctuation duration refers to the cumulative duration that the anomaly triggering node's performance indicator exceeds the preset threshold, which is used to illustrate the continued impact of the anomaly. The abnormal characteristic parameters refer to the key parameter set that can characterize the abnormal state of the anomaly triggering node, here namely the fluctuation frequency, fluctuation amplitude, and fluctuation duration. The anomaly analysis report is a document that integrates information such as abnormal characteristic parameters, abnormal propagation links, and the scope of abnormal impact to present the root cause, development process, and impact of the anomaly. For example, if the fluctuation frequency of an anomaly triggering node is 3 times, the fluctuation amplitude is 15%, and the fluctuation duration is 20 minutes, these parameters will be included in the report to illustrate the abnormal characteristics of the node.

[0099] The specific implementation process of this step is as follows: first, organize the fluctuation frequency, fluctuation amplitude and fluctuation duration obtained in S209, and clarify the specific values ​​and units of each parameter (such as frequency unit is "times", amplitude unit is "%", and time unit is "minutes"); then, combine the abnormal propagation link (such as "trigger node → object A → object B"), the abnormal impact of the target abnormal object, the performance change characteristics of the associated objects and other information to build the core content framework of the report, including the abnormal trigger node identification, abnormal characteristic parameter details, abnormal propagation path, list of affected objects and the degree of impact; then, structure this information, for example, present "abnormal source analysis", "propagation process", "impact assessment" and other contents in chapters, among which the abnormal characteristic parameter part needs to list the specific values ​​of frequency, amplitude and duration in detail, and attach a performance fluctuation curve chart for auxiliary explanation; finally, generate a standardized document (such as PDF format), namely the abnormal analysis report, to provide operation and maintenance personnel with a clear basis for fault diagnosis.

[0100] In actual operation, reports can be automatically generated through programming: for example, using Python's ReportLab library to read abnormal characteristic parameter data, insert text and charts according to preset templates (such as fluctuation trend charts drawn with Matplotlib), automatically calculate the quantitative score of the abnormal impact (such as weighted calculation based on frequency, amplitude, and duration), and finally output a report file containing complete analysis content to ensure that the information is accurate and easy to understand.

[0101] The platform operation performance evaluation system in the embodiment of the present invention is described below from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of the platform operation performance evaluation system in an embodiment of the present application.

[0102] It should be noted that Figure 3The structure of the platform operation performance evaluation system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0103] like Figure 3 As shown, the platform operation performance evaluation system includes a CPU (Central Processing Unit) 301, which can perform various appropriate actions and processes based on programs stored in ROM (Read-Only Memory) 302 or programs loaded from storage 308 into RAM (Random Access Memory) 303, such as executing the methods described in the above embodiments. RAM 303 also stores various programs and data required for system operation. CPU 301, ROM 302, and RAM 303 are interconnected via bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0104] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, push button switches, and the like; an output section 307 including a liquid crystal display (LCD), an audio output device, indicator lights, and the like; a storage section 308 including a hard disk and the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 310 as needed, so that computer programs read from the removable media can be installed in the storage section 308 as needed.

[0105] In particular, according to an embodiment of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 309 and / or installed from the removable medium 311. When the computer program is executed by the CPU 301, the various functions defined in the present invention are performed.

[0106] It should be noted that specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0107] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings.

[0108] Specifically, the platform operation performance evaluation system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the platform operation performance evaluation method based on collaborative intelligent analysis provided in the above embodiment is implemented.

[0109] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the platform operation performance evaluation system described in the above embodiments, or may exist independently and not be incorporated into the platform operation performance evaluation system. The storage medium carries one or more computer programs, and when the one or more computer programs are executed by a processor of the platform operation performance evaluation system, the platform operation performance evaluation system implements the platform operation performance evaluation method based on collaborative intelligent analysis provided in the above embodiments.

[0110] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0111] As used in the above embodiments, the term “when” may be interpreted to mean “if” or “after” or “in response to determining that” or “in response to detecting that”, depending on the context. Similarly, the phrases “upon determining that” or “if (stated condition or event) is detected” may be interpreted to mean “if determining that” or “in response to determining that” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0112] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A platform operation performance evaluation method based on collaborative intelligent analysis, characterized in that: Applied to a platform operation performance evaluation system, the method includes: Acquire data association relationships between monitoring objects in the platform to be evaluated, and establish a directed data dependency graph based on the data association relationships, wherein each edge in the directed data dependency graph represents a data interaction relationship between the monitoring objects; Collecting the operating performance data of the monitored object, and arranging the operating performance data in chronological order to form a performance fluctuation sequence; Calculating the association importance of the monitoring object in the directed data dependency graph, wherein the association importance is weightedly calculated according to the number of direct associations and the number of indirect associations of the monitoring object; Analyze the performance fluctuation sequence, calculate a propagation consistency factor of performance data exceeding a preset performance threshold, and a temporal closeness factor of associated performance data in the performance fluctuation sequence, wherein the propagation consistency factor indicates the degree of propagation of the performance data between associated objects, and the temporal closeness factor indicates the degree of temporal correlation of the associated performance data; Calculating a propagation weight value between any two of the monitored objects and constructing a performance fluctuation propagation sequence of the monitored objects based on the propagation weight value, wherein the propagation weight value is obtained by weighted calculation of the association importance, the propagation consistency factor, and the time closeness factor, and the propagation sequence represents the transmission order of performance anomalies between the monitored objects; According to the performance fluctuation propagation sequence and the directed data dependency graph, the abnormal impact of the monitored object is calculated, and the monitored object whose abnormal impact exceeds the preset impact threshold is determined as the target abnormal object. The abnormal impact is calculated based on the position of the monitored object in the propagation sequence, the number of downstream objects affected, and the associated importance of each object.

2. The method according to claim 1, characterized in that The step of collecting the operating performance data of the monitored object and arranging the operating performance data in chronological order to form a performance fluctuation sequence specifically includes: Collecting performance indicator data including the monitored object, the performance indicator data including at least processor usage data and memory occupancy data; Obtaining sampling values ​​of the performance indicator data according to a preset sampling time interval; The sampling value is associated with the corresponding sampling time to generate performance indicator time series data, and the performance fluctuation sequence is constructed according to the performance indicator time series data, wherein each sequence element includes a time identifier and corresponding performance indicator data.

3. The method according to claim 1, characterized in that The step of calculating the association importance of the monitoring object in the directed data dependency graph, wherein the association importance is weightedly calculated according to the number of direct associations and the number of indirect associations of the monitoring object, specifically includes: Counting the number of adjacent nodes of the monitored object in the directed data dependency graph, and using the number of adjacent nodes as the number of directly associated nodes; Traversing the directed data dependency graph, obtaining the number of all reachable nodes of the monitored object, and using the number of reachable nodes as the number of indirect associations; Multiplying the number of direct associations by a first preset weight to obtain a direct association score, and multiplying the number of indirect associations by a second preset weight to obtain an indirect association score; The direct correlation score and the indirect correlation score are added together to obtain the correlation importance of the monitored object.

4. The method according to claim 1, wherein The step of calculating the abnormal impact of the monitored object according to the performance fluctuation propagation sequence and the directed data dependency graph, and determining the monitored object whose abnormal impact exceeds a preset impact threshold as a target abnormal object specifically includes: Obtaining a sequence number of the monitored object in the performance fluctuation propagation sequence, and determining a position weight value of the monitored object according to a preset position weight calculation rule; Determine the number of all downstream nodes of the monitored object in the directed data dependency graph, and multiply the position weight value by the number of downstream nodes to obtain an initial influence value; Calculating the sum of the associated importances of the downstream nodes, and multiplying the initial impact value by the sum of the associated importances to obtain the abnormal impact of the monitored object; It is determined whether the abnormal impact is greater than the preset impact threshold; if so, the monitored object is determined as the target abnormal object.

5. The method according to claim 1, characterized in that After the step of calculating the abnormal impact of the monitored object based on the performance fluctuation propagation sequence and the directed data dependency graph, and determining the monitored object whose abnormal impact exceeds a preset impact threshold as a target abnormal object, the method further includes: Obtaining abnormal points of performance indicators based on the performance fluctuation sequence of the target abnormal object; In the directed data dependency graph, starting from the target abnormal object, the performance fluctuation data of the associated objects of the target abnormal object are obtained according to a preset hierarchy; Performing time series alignment on the performance fluctuation data to determine the time series relationship of the performance anomaly; Determining an abnormal propagation link based on the time sequence relationship and the topological structure of the directed data dependency graph; On the abnormal propagation link, an abnormal triggering node is identified, and abnormal characteristic parameters of the abnormal triggering node are counted to generate an abnormal analysis report.

6. The method according to claim 5, characterized in that The step of determining the abnormal propagation link according to the time sequence relationship and in combination with the topological structure of the directed data dependency graph specifically includes: Obtaining a fluctuation trend of the performance indicator, and determining an analysis time window according to the fluctuation trend; Extracting performance indicator change characteristics of objects associated with the target abnormal object within the analysis time window; Constructing a propagation path of performance fluctuations based on the topological structure of the directed data dependency graph and the performance indicator change characteristics; The propagation order of the associated objects in the propagation path of the performance fluctuation is calculated, and the abnormal propagation link is determined according to the propagation order.

7. The method according to claim 5, characterized in that The step of identifying an abnormality triggering node on the abnormality propagation link and generating an abnormality analysis report by counting abnormal characteristic parameters of the abnormality triggering node specifically includes: Determining, according to the propagation direction of the abnormal propagation link, a starting node of the abnormal propagation link as the abnormality triggering node; Extracting the performance fluctuation characteristics of the abnormal triggering node, and counting the fluctuation frequency, fluctuation amplitude and fluctuation duration of the abnormal triggering node according to the performance fluctuation characteristics; The fluctuation frequency, the fluctuation amplitude and the fluctuation duration are used as the abnormal characteristic parameters, and the abnormal analysis report is generated according to the abnormal characteristic parameters.

8. A platform operation performance evaluation system, characterized in that: The platform operation performance evaluation system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the platform operation performance evaluation system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a platform operation performance evaluation system, the platform operation performance evaluation system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a platform operation performance evaluation system, the platform operation performance evaluation system is caused to execute the method according to any one of claims 1 to 7.

Images