Multi-dimensional performance bottleneck diagnosis method and device based on call chain, medium and product

Abstract

The invention discloses a multi-dimensional performance bottleneck diagnosis method and device based on a call chain, a medium and a product, and the method comprises the steps: collecting call chain data in a target service system, splitting the call chain data into processing unit data, and recognizing target unit data meeting a high-time-consumption condition; constructing a dimension list of the target service system according to the system attribute of the target service system; calculating dimension importance according to the time consumption value of the target unit data under each dimension value, and screening a set number of target dimensions; combining the dimension values of the at least two target dimensions to obtain combined dimension values, and calculating an interaction effect value according to the time consumption value of the target unit data under each combined dimension value; according to the technical scheme provided by the embodiment of the invention, the calculation complexity is effectively reduced through an intelligent dimension screening mechanism, the problem of dimension explosion is avoided, the analysis efficiency is improved, the root dimension combination of the performance bottleneck is accurately positioned, and the accuracy and operability of the diagnosis result are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a call chain-based multi-dimensional performance bottleneck diagnosis method, device, medium and product. Background Art

[0002] With the rapid popularization of modern distributed system architecture and the widespread application of microservice architecture, the system topology has become more complex, which makes performance bottleneck diagnosis face significant challenges.

[0003] In existing technologies, performance bottleneck diagnosis mainly adopts two methods: basic indicator monitoring collects computing core, memory and other resource usage data and sets alarm thresholds for monitoring; call chain tracing technology is based on standards such as OpenTelemetry (Open Link Tracing), records call chain data containing TraceID (TraceIdentifier, tracing identifier), service node and time consumption, and constructs a service call topology diagram.

[0004] In the process of implementing the present invention, the inventors discovered that although traditional basic indicator monitoring can identify single-point resource anomalies, it cannot track the transmission path of cross-service performance problems, resulting in one-sided diagnostic results; and although simple call chain analysis can display the service call topology, it lacks multi-dimensional intelligent correlation analysis capabilities, making it difficult to accurately locate the root cause of performance bottlenecks. Summary of the Invention

[0005] The embodiments of the present invention provide a multi-dimensional performance bottleneck diagnosis method, device, medium and product based on call chain, which can realize intelligent and accurate diagnosis of distributed system performance bottlenecks.

[0006] According to one aspect of an embodiment of the present invention, a multi-dimensional performance bottleneck diagnosis method based on a call chain is provided, the method comprising:

[0007] Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data;

[0008] The target service system includes multiple service nodes, which are set in a set area and are used to implement processing functions that match the set processing units in the call chain;

[0009] Constructing a dimension list of the target service system according to the system properties of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values;

[0010] Calculate the dimension importance of each dimension based on the processing time of the target unit data of each dimension value under each dimension, and filter out a set number of target dimensions based on the dimension importance;

[0011] Combining the dimensional values ​​of at least two target dimensions to obtain multiple combined dimensional values, and calculating the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value;

[0012] According to the interaction effect value of each combined dimension value, the multi-dimensional performance bottleneck diagnosis result of the target service system is generated.

[0013] According to another aspect of an embodiment of the present invention, a multi-dimensional performance bottleneck diagnosis device based on a call chain is provided, the device comprising:

[0014] A data collection module is configured to collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high-time-consuming condition in each processing unit data; wherein the target service system includes multiple service nodes, which are arranged in a set region and are configured to implement processing functions that match the set processing units in the call chain;

[0015] A multi-dimensional system module is used to construct a dimension list of the target service system according to the system properties of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values;

[0016] The dimension screening module is used to calculate the dimension importance of each dimension based on the processing time value of the target unit data of each dimension value under each dimension, and screen out a set number of target dimensions based on the dimension importance;

[0017] A combination analysis module is used to combine the dimensional values ​​of at least two target dimensions to obtain multiple combined dimensional values, and calculate the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value;

[0018] The diagnosis decision module is used to generate multi-dimensional performance bottleneck diagnosis results of the target service system based on the interaction effect value of each combined dimension value.

[0019] According to another aspect of an embodiment of the present invention, an electronic device is provided, the electronic device comprising:

[0020] at least one processor; and

[0021] a memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the multi-dimensional performance bottleneck diagnosis method based on call chain described in any embodiment of the present invention.

[0023] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement a multi-dimensional performance bottleneck diagnosis method based on a call chain as described in any embodiment of the present invention when executed.

[0024] According to another aspect of an embodiment of the present invention, a computer program product is provided, comprising computer instructions, which implement the steps of the method according to any embodiment of the present invention when executed by a processor.

[0025] The technical solution of an embodiment of the present invention collects call chain data in a target service and splits each call chain data into processing unit data, identifies target unit data that meets high time-consuming conditions in the processing unit data, wherein the target service includes multiple service nodes set in a set area, and these nodes are used to implement processing functions that match the set processing units in the call chain. Subsequently, a dimension list containing multiple dimensions is constructed according to the attributes of the target service, and each dimension contains multiple dimension values. The dimension importance of each dimension is calculated based on the processing time value of the target unit data under each dimension value, and a set number of target dimensions are screened out accordingly. Then, the dimension values ​​under at least two target dimensions are combined to obtain multiple combined dimension values, and the interaction effect value of each combined dimension value is calculated based on the processing time value of the target unit data under each combined dimension value. Finally, a multi-dimensional performance bottleneck diagnosis result is generated based on these interaction effect values. This new performance bottleneck diagnosis method effectively solves the problems of dimensional explosion and rough analysis faced by traditional methods in complex scenarios by establishing a systematic analysis process, providing a new technical path for performance optimization of distributed systems, significantly improving diagnosis efficiency while ensuring analysis accuracy.

[0026] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 This is a flowchart of a multi-dimensional performance bottleneck diagnosis method based on call chains according to the first embodiment of the present invention;

[0029] Figure 2 This is a flowchart of another method for diagnosing multi-dimensional performance bottlenecks based on call chains according to the second embodiment of the present invention;

[0030] Figure 3 This is a flowchart of another method for diagnosing multi-dimensional performance bottlenecks based on call chains according to the third embodiment of the present invention;

[0031] Figure 4 2 is a schematic diagram of the structure of a multi-dimensional performance bottleneck diagnosis device based on call chain according to the fourth embodiment of the present invention;

[0032] Figure 5 It is a structural diagram of an electronic device that implements a multi-dimensional performance bottleneck diagnosis method based on a call chain according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] Example 1

[0036] Figure 1This is a flowchart of a multi-dimensional performance bottleneck diagnosis method based on a call chain provided in Example 1 of the present invention. This embodiment is applicable to the performance bottleneck diagnosis of a distributed system under a microservice architecture. The method can be executed by a multi-dimensional performance bottleneck diagnosis device based on a call chain. The device can be implemented in the form of hardware and / or software and can generally be configured in an electronic device.

[0037] Correspondingly, such as Figure 1 As shown, the method includes:

[0038] S110: Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data;

[0039] The target service system includes a plurality of service nodes, which are arranged in a set region and used to implement processing functions that match the set processing units in the call chain.

[0040] Among them, call chain data can be understood as the core monitoring data that records the execution trajectory of the entire request chain in a distributed system. The call relationship across service nodes is connected into a complete topology through a unique TraceID. This type of data strictly follows the OpenTelemetry standard specification and contains key fields such as service node name, interface path, and time value. Processing unit data can be understood as the basic execution unit (span) in the call chain. Each span corresponds to a complete operation record of a specific service node, including core information such as execution time and interface path. To maintain terminology consistency, this article will uniformly use "span" to refer to this concept.

[0041] High-cost conditions can be understood as statistically defined performance anomaly criteria. This is achieved by calculating the percentile value of the duration distribution of all spans in the call chain as a baseline threshold. For example, the 95th percentile is chosen as the baseline threshold for high-cost execution, denoted by P95. When a span's execution time exceeds this threshold, it is marked as a high-cost target unit. A service node can be understood as the smallest service unit that performs a specific function in a distributed system. Each node corresponds to an independently deployed service instance with a clear physical location and service identity. These nodes call each other through standard interfaces and are reflected as independent span records in the call chain data. Their operating status and performance metrics form the fundamental unit for system performance analysis.

[0042] In this example, call chain data is broken down into independent spans using a standardized collection process. High-cost target units are then filtered based on preset percentile thresholds. This data originates from service nodes deployed within a defined region. Each node is responsible for executing a specific processing function in the call chain, providing atomic-level performance observation units for subsequent analysis.

[0043] S120: Construct a dimension list of the target service system according to the system attributes of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values.

[0044] The dimension list can be understood as a multi-dimensional analysis framework built based on the target service attributes. The list contains multiple independent dimensions (such as service name and region), and each dimension is further subdivided into specific dimension values ​​(such as service1 / service2 or region-east / region-west).

[0045] In this embodiment, the essence of the dimension list is to convert service attributes (such as service name, interface path) and operating environment characteristics (such as deployment region, instance version) into standardized analysis dimensions. These dimensions form a complete classification system through hierarchical organization, and each dimension value accurately corresponds to a specific service operation status. In the actual analysis process, the dimension list plays a key role: first, as a data association hub, it associates raw performance data with business characteristics through span tags; second, it provides a basic framework for computing optimization and supports core algorithms such as information gain calculation and interaction effect analysis. For example, basic dimensions such as service name and region can directly locate the scope of the anomaly, while dynamically derived dimensions such as error type and traffic level can deeply analyze the nature of the problem.

[0046] S130 , calculating the dimension importance of each dimension according to the processing time consumption value of the target unit data of each dimension value under each dimension, and screening out a set number of target dimensions according to the dimension importance.

[0047] Dimension importance can be understood as a key metric for evaluating the impact of different dimensions on system performance. It reflects the ability of a specific dimension to discriminate between time-consuming anomalies. This metric measures the importance of a dimension in performance analysis by quantifying the degree to which the performance of each dimension value (such as service1, region-east) differs from the overall baseline. The higher the dimension importance, the greater its role in identifying and locating performance bottlenecks, helping operations personnel more effectively focus on key issues.

[0048] In this embodiment, the importance score for each dimension is calculated based on the data processing time of the high-time target unit corresponding to each dimension value in the dimension list. Through a preset screening mechanism, a number of target dimensions whose importance meets the threshold requirement are selected from all dimensions. These target dimensions will serve as the basis for subsequent multi-dimensional combined analysis.

[0049] S140. Combine the dimensional values ​​under at least two target dimensions to obtain multiple combined dimensional values, and calculate the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value.

[0050] Among them, the combined dimension can be understood as a composite analysis unit formed by cross-combining the dimension values ​​under multiple target dimensions. For example, the two dimension values ​​of "service=service1" and "region=us-east" are combined to form a new analysis dimension combination. This combination method can reveal complex performance characteristics that cannot be displayed by a single dimension and is the basis for analyzing multi-dimensional correlation effects. The interaction effect value can be understood as a value used to quantify the degree of synergistic influence generated by the combination of multiple dimensions, reflecting the additional impact of the interaction between dimensions on performance. This indicator can identify special effects that exceed the sum of the independent effects of each dimension, such as abnormal performance fluctuations of a specific service in a specific region, which cannot be explained by analyzing the service or region dimensions alone.

[0051] In this embodiment, the selected target dimensions are combined in pairs or multiple combinations to generate combined dimensions containing multiple dimension values. Based on the processing time of the high-time target unit data corresponding to each combined dimension, an interaction effect value is calculated to reflect the synergistic influence between the dimensions, providing data support for identifying abnormal patterns in complex scenarios.

[0052] S150 : Generate a multi-dimensional performance bottleneck diagnosis result of the target service system according to the interaction effect value of each combined dimension value.

[0053] In this embodiment, based on the calculated interaction effect values ​​of each combination dimension, sorting and screening are performed according to the size of the values. Combinations with high interaction effects indicate high abnormal correlation, thereby identifying abnormal patterns. For example, service cascade abnormality patterns, when the service is abnormal, the order and billing services are also abnormal. Regional abnormal patterns can also be discovered, such as performance degradation in the Beijing area.

[0054] The technical solution of an embodiment of the present invention collects call chain data in a target service and splits each call chain data into processing unit data, identifies target unit data that meets high time-consuming conditions in the processing unit data, wherein the target service includes multiple service nodes set in a set area, and these nodes are used to implement processing functions that match the set processing units in the call chain. Subsequently, a dimension list containing multiple dimensions is constructed according to the attributes of the target service, and each dimension contains multiple dimension values. The dimension importance of each dimension is calculated based on the processing time value of the target unit data under each dimension value, and a set number of target dimensions are screened out accordingly. Then, the dimension values ​​under at least two target dimensions are combined to obtain multiple combined dimension values, and the interaction effect value of each combined dimension value is calculated based on the processing time value of the target unit data under each combined dimension value. Finally, a multi-dimensional performance bottleneck diagnosis result is generated based on these interaction effect values. This new performance bottleneck diagnosis method effectively solves the problems of dimensional explosion and rough analysis faced by traditional methods in complex scenarios by establishing a systematic analysis process, providing a new technical path for performance optimization of distributed systems, significantly improving diagnosis efficiency while ensuring analysis accuracy.

[0055] Example 2

[0056] Figure 2 This is a flowchart of another method for diagnosing multidimensional performance bottlenecks based on call chains, provided in Example 2 of the present invention. This example optimizes the above examples. Specifically, the operation of "calculating the dimensional importance of each dimension based on the processing time of the target unit data for each dimensional value under each dimension" is refined.

[0057] Correspondingly, such as Figure 2 As shown, the method includes:

[0058] S210: Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data;

[0059] The target service system includes a plurality of service nodes, which are arranged in a set region and used to implement processing functions that match the set processing units in the call chain.

[0060] S220: Construct a dimension list of the target service system according to the system attributes of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values.

[0061] S230 , obtaining the current dimension in the dimension list, and obtaining the proportion of the target unit data in each processing unit data belonging to the current dimension as the baseline high time consumption rate.

[0062] The baseline high-cost rate can be understood as the proportion of high-cost target unit data among all processing unit data in the current dimension of the target service system. It reflects the overall high-cost situation in that dimension and serves as a benchmark for subsequent comparisons. For example, if there are 1000 spans in a dimension and 150 of them are high-cost data, the baseline high-cost rate is 15%.

[0063] In this embodiment, for the current dimension in the dimension list, the proportion of all processing unit data under this dimension that is identified as target unit data is counted to calculate the baseline high time consumption rate of this dimension. This indicator reflects the frequency of high time consumption within the entire scope of the current dimension and provides a benchmark reference value for subsequent dimension importance calculations. The formula for the baseline high time consumption rate is as follows:

[0064]

[0065] Here, is_high_latency(span) represents the high-latency span of the current dimension, and |D| represents the total number of spans in the current dimension.

[0066] S240 , respectively obtaining the proportion of the target unit data in each processing unit data of each current dimension value under the current dimension as the segment high time consumption rate of each current dimension value.

[0067] In this embodiment, for each specific dimension value within the current dimension, the percentage of target unit data identified within the corresponding processing unit data for that dimension value is counted, and the high time consumption rate for each dimension value is calculated. This metric reflects the frequency of high time consumption for each specific dimension value (such as service1 and service2). By comparing it with the overall baseline high time consumption rate for the dimension, it can accurately identify specific dimension values ​​that are performing abnormally.

[0068] S250: Calculate the dimension importance of the current dimension based on the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each current dimension value under the current dimension.

[0069] Dimension importance can be understood as a key indicator that measures the impact of a specific dimension on overall performance anomalies. Its core reflects the degree of deviation between the segment high-consumption rate of each dimension value under that dimension and the baseline high-consumption rate. For example, when the segment high-consumption rate of some dimension values ​​(such as "service = service 1") under a dimension (such as "service") is significantly higher or lower than the dimension baseline, the dimension is given a higher importance, indicating that it can effectively distinguish performance anomaly scenarios.

[0070] In this embodiment, a dimension importance score is calculated based on the difference between the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each dimension value using a preset calculation method. This score quantitatively reflects the degree of dispersion of the performance of each dimension value under the current dimension. A larger score indicates that the dimension is more valuable for identifying high time consumption anomalies.

[0071] S260. Combine the dimensional values ​​under at least two target dimensions to obtain multiple combined dimensional values, and calculate the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value.

[0072] S270. Generate a multi-dimensional performance bottleneck diagnosis result of the target service system according to the interaction effect value of each combined dimension value.

[0073] The technical solution of the embodiment of the present invention collects the call chain data of the target service, splits it into processing unit data and identifies high-time-consuming target unit data. These data come from service nodes that perform specific processing functions in a set area. Then, a dimension list containing multiple dimensions is constructed according to the attributes of the target service. Each dimension contains multiple dimension values. For the current dimension in the dimension list, the proportion of the number of target unit data identified as target unit data in all processing unit data under the dimension is counted as the baseline high-time-consuming rate. At the same time, the proportion of the number of target unit data in each dimension value under the current dimension is counted as the segment high-time-consuming rate. According to the current dimension The dimension importance of the current dimension is calculated by combining the baseline high time consumption rate and the segment high time consumption rate of each dimension value, and a set number of target dimensions are screened out. The screened target dimensions are combined, and the interaction effect value is calculated based on the high time consumption data under the combined dimension value, and finally a multi-dimensional performance bottleneck diagnosis result is generated. This new performance bottleneck diagnosis method can intelligently identify the key dimensions that have the greatest impact on system performance through the established dimension importance evaluation mechanism, effectively avoiding the subjectivity and one-sidedness of traditional manual experience analysis; at the same time, the baseline high time consumption rate and segment high time consumption rate comparison method based on statistical principles ensures the objectivity and accuracy of the dimension screening results.

[0074] Furthermore, based on the above embodiments, identifying target unit data that meets the high time consumption condition in each processing unit data may further include:

[0075] Obtaining the processing time value corresponding to each processing unit data, and filtering the processing unit data whose processing time value is outside a preset reasonable time range;

[0076] Identify target unit data that meets the high time consumption condition in each processing unit data, including:

[0077] Calculate the percentile of the target data indicator based on the processing time of each processing unit data;

[0078] The processing unit data whose processing time value exceeds the percentile under the target data indicator is extracted as the target unit data that meets the high time consumption condition.

[0079] Generally speaking, performance analysis begins with preprocessing the collected raw call chain data. This involves obtaining the processing time corresponding to each span and filtering out data outside a preset reasonable range, such as abnormal records with a processing time less than 0 or exceeding a maximum threshold (for example, a threshold set to 60 seconds), to ensure the accuracy of subsequent analysis.

[0080] Generally, quantile statistics are used to identify high-cost spans. Specifically, the duration of all spans is first extracted and sorted from smallest to largest. Then, criteria for determining high-cost spans are set based on business requirements. For example, spans with durations ranked at or above the 95th percentile are considered high-cost target units. This means that these spans exceed 95% of other spans and are among the worst-performing 5% of requests. This sorting-based quantile determination method objectively reflects the relative position of a span in the overall time distribution.

[0081] Generally speaking, this quantile screening mechanism has the advantage of dynamic adaptability compared to fixed thresholds. Since the judgment criteria are based on the relative position of the actual time consumption distribution rather than the absolute value, it can automatically adapt to changes in performance characteristics under different business scenarios and system loads.

[0082] Optionally, based on the above embodiments, the dimension importance of the current dimension is calculated according to the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each current dimension value under the current dimension, which may include:

[0083] Calculate the difference between the segment high time consumption rate and the baseline high time consumption rate of each current dimension value and multiply it by the number of processing unit data belonging to each current dimension value to obtain the information gain contribution value of each current dimension value;

[0084] The information gain contribution values ​​of each current dimension value are accumulated and summed to obtain the dimension importance of the current dimension.

[0085] Generally, dimension importance is calculated using an information gain algorithm to quantify each dimension's ability to discriminate against performance anomalies. For each dimension value, the absolute difference between its segment's high-consumption rate and the dimension's overall baseline high-consumption rate is calculated. This difference reflects the degree of performance deviation for that dimension value. This difference is then multiplied by the number of processing unit data points corresponding to that dimension value to obtain the information gain contribution for that dimension value. This product takes into account both the anomaly intensity and the impact range.

[0086] Generally speaking, the final importance of a dimension is obtained by aggregating the contribution values ​​of all the dimension values ​​under it. Specifically, the information gain contribution values ​​of all dimension values ​​under the current dimension are accumulated and summed. This accumulated sum is the dimension importance score of the dimension. The dimension importance calculation formula is as follows:

[0087]

[0088] Among them, v∈values(dim) represents the set of all dimension values ​​of dimension dim, rate(v) represents the segment high time consumption rate of dimension value v, baselin_rate represents the baseline high time consumption rate of dimension dim, and count(v) is the number of processing unit data belonging to dimension value v.

[0089] Example 3

[0090] Figure 3 This is a flowchart of another method for diagnosing multidimensional performance bottlenecks based on call chains, provided in Example 2 of the present invention. This example is optimized based on the above examples. Specifically, the operation of "calculating the interaction effect value of each combined dimension value based on the processing time value of the target unit data belonging to each combined dimension value" is refined.

[0091] Correspondingly, such as Figure 3 As shown, the method includes:

[0092] S310: Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data;

[0093] The target service system includes a plurality of service nodes, which are arranged in a set region and used to implement processing functions that match the set processing units in the call chain.

[0094] S320: Construct a dimension list of the target service system according to the system attributes of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values.

[0095] S330 , calculating the dimension importance of each dimension according to the processing time consumption value of the target unit data of each dimension value under each dimension, and screening out a set number of target dimensions according to the dimension importance.

[0096] S340: Combine the dimensional values ​​of at least two target dimensions to obtain multiple combined dimensional values.

[0097] S350 , respectively obtaining the quantity value of the target unit data and the proportion of the target unit data in each processing unit data belonging to each combined dimension value.

[0098] In this example, due to the complex interactions between the dimensions, the combined duration is not simply a linear summation, but rather produces unpredictable nonlinear variations. To accurately assess this multi-dimensional interaction, a dual statistical approach was performed for each combined dimension: first, the number of spans identified as time-consuming spans in that combination was counted to reflect the scale of the problem; and second, the proportion of these time-consuming spans in the total spans of that combination was calculated to measure the severity of the anomaly.

[0099] S360. Calculate the influence ratio corresponding to each combined dimension value according to the quantity value of the target unit data corresponding to each combined dimension value and the proportion of the target unit data.

[0100] The influence ratio can be understood as a comprehensive indicator that quantifies the contribution of a dimension combination to overall performance issues. Its calculation considers two key factors: the number of time-consuming spans associated with that dimension combination and its proportion of total spans. By applying a standardized weighted calculation to these two dimensions, the ratio objectively assesses the actual impact of each dimension combination on system performance bottlenecks. A larger value indicates a more significant contribution to the overall performance issue.

[0101] In this embodiment, the influence ratio of the combined dimension value is calculated by a preset algorithm based on the number of high-time-consuming target unit data corresponding to each combined dimension value and its proportion. This calculation process strictly follows the actual statistical scale of high-time-consuming data and the abnormal proportion under the combined dimension value, and standardizes the quantity value and the proportion index according to the established rules to generate a quantitative evaluation result that can be compared horizontally. The value of the influence ratio directly reflects the relative contribution of the dimension combination to the overall performance problem, providing an accurate ranking basis for subsequent cumulative screening. The influence ratio formula is as follows:

[0102] Impact ratio = (number of high-time-consuming requests × high-time-consuming ratio) / 100

[0103] S370. In descending order of influence ratio, the influence ratios of the combined dimension values ​​are successively added up. When the sum is greater than or equal to the target ratio value, the combined dimension values ​​currently involved in the calculation are used as the target combined dimension values.

[0104] By gradually accumulating high-impact dimension combinations, we screen out the key combinations that contribute most to performance issues. We use a high-to-low ranking and accumulation mechanism, stopping when the cumulative impact reaches a preset target ratio (typically set at 50%). This ensures that the selected combinations cover the vast majority of performance anomalies while avoiding analyzing too many low-value combinations.

[0105] S380. Calculate the interaction effect value of each target combination dimension value based on the proportion of target unit data corresponding to each target combination dimension value and the expected proportion of the target unit data.

[0106] The expected proportion of target unit data corresponding to the target combination dimension value is calculated by calculating the baseline high time consumption rate of each dimension in the target combination dimension value.

[0107] In this example, the potential interaction effects between dimension combinations are revealed by comparing the differences between actual observed values ​​and theoretical expected values. Specifically, the expected proportion is first calculated based on the baseline high time consumption rate of each component dimension, and then compared with the actual observed high time consumption proportion. The degree of difference between the two is the interaction effect value. A positive value indicates that the combination has a synergistic effect that exacerbates performance issues, a negative value suggests an inhibitory effect, and a zero value indicates no significant interaction. The interaction effect formula is as follows:

[0108] Interaction effect = high time consumption ratio of dimension combination - expected high time consumption ratio

[0109] in, N is the number of dimensions.

[0110] S390. Generate a multi-dimensional performance bottleneck diagnosis result of the target service system based on the interaction effect value of each combined dimension value.

[0111] The technical solution of the embodiment of the present invention collects call chain data of the target service and splits it into processing unit data, thereby identifying high-time-consuming target unit data. This data comes from service nodes that perform specific processing functions within a set region. Subsequently, a dimension list containing multiple dimensions and dimension values ​​is constructed based on service attributes. The baseline high-time consumption rate of each dimension and the segment high-time consumption rate of each dimension value are calculated. Based on this, the dimension importance is evaluated and key target dimensions are screened. The dimension values ​​of the target dimensions are combined, and the number and proportion of high-time-consuming data in each combination are counted. The influence ratio is calculated, and the ratios are accumulated from large to small to the target ratio value. The core target combination dimension values ​​are screened. By comparing the actual proportion with the expected proportion based on the baseline rate of each dimension, the interaction effect value is calculated, and finally a multi-dimensional performance bottleneck diagnosis result is generated. This new performance bottleneck diagnosis method can intelligently identify the key dimension combinations that have the greatest impact on system performance through a quantitative screening mechanism, effectively avoiding the subjectivity and one-sidedness of traditional manual experience analysis. At the same time, the influence ratio accumulation method based on statistical principles ensures the objectivity and accuracy of the dimension combination screening results, significantly improving the efficiency and accuracy of performance bottleneck diagnosis in complex distributed environments.

[0112] Optionally, based on the above embodiments, combining the dimension values ​​under at least two target dimensions to obtain multiple combined dimension values ​​may include:

[0113] Combining the target dimensions according to a preset number of optional combinations to obtain at least one dimension set, wherein each dimension set contains at least two target dimensions;

[0114] Combining the dimensional values ​​of each target dimension in each dimensional set with the dimensional values ​​of other target dimensions in the dimensional set to obtain multiple combined candidate dimensional values;

[0115] The quantity value of each processing unit data belonging to each combined candidate dimension value is calculated respectively, and the combined candidate dimension value whose quantity value is greater than or equal to the preset quantity threshold is determined as the combined dimension value.

[0116] Generally speaking, after selecting the target dimensions according to the importance of each dimension, these target dimensions are arranged and combined according to the preset combination rules (such as pairwise combination or multi-dimensional combination) to generate a dimension set containing different dimension combinations, where each dimension set consists of at least two target dimensions; then, within each dimension set, the dimension values ​​of all target dimensions will be fully arranged and combined to generate a complete dimension value combination scheme.

[0117] Specifically, in a specific example: by sorting the dimensions by importance, the four most important dimensions are retained: service, interface, region and ip. Based on the four important dimensions, different levels of combinations are generated, of which there are 6 binary combinations (service×interface, service×region, service×ip, region×interface, ip×interface, region×ip) and 4 ternary combinations (service×interface×region, service×interface×ip, service×region×ip, interface×region×ip).

[0118] Generally, the processing unit data for each candidate dimension combination is counted and filtered using a preset threshold. Only when the processing unit data for a candidate dimension combination reaches or exceeds the threshold is it retained as a valid dimension combination. Typically, the threshold is set to 100. This mechanism effectively filters out statistical noise caused by insufficient data and improves the reliability of subsequent analysis.

[0119] Optionally, based on the above embodiments, generating a multi-dimensional performance bottleneck diagnosis result of the target service system according to the interaction effect value of each combined dimension value may include:

[0120] identifying at least one abnormal pattern in each target combination dimension value based on the interaction effect value of each target combination dimension value;

[0121] Based on the identified abnormal patterns, multi-dimensional performance bottleneck diagnosis results of the target service system are generated.

[0122] Generally speaking, the interaction effect value quantifies the synergistic impact of a dimension combination on a performance issue. A positive value indicates that the combination exacerbates the performance issue, while a negative value suggests a mitigating effect. By setting an appropriate threshold range, we can screen for dimension combinations with significant positive interaction effects. These combinations are identified as abnormal patterns that lead to performance bottlenecks. In practice, we typically select the top few combinations with interaction effect values ​​exceeding the preset threshold as key abnormal patterns.

[0123] Example 4

[0124] Figure 4 This is a schematic diagram of the structure of a multi-dimensional performance bottleneck diagnosis device based on call chain provided by the fourth embodiment of the present invention. Figure 4 As shown, the device includes:

[0125] Data collection module 410 is configured to collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high-time-consuming condition in each processing unit data. The target service system includes multiple service nodes, which are arranged in a predetermined region and are configured to implement processing functions that match the predetermined processing units in the call chain.

[0126] The multi-dimensional system module 420 is used to construct a dimension list of the target service system according to the system attributes of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values;

[0127] The dimension screening module 430 is used to calculate the dimension importance of each dimension based on the processing time of the target unit data of each dimension value under each dimension, and screen out a set number of target dimensions based on the dimension importance;

[0128] A combination analysis module 440 is configured to combine the dimensional values ​​of at least two target dimensions to obtain a plurality of combined dimensional values, and calculate the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value;

[0129] The diagnosis decision module 450 is used to generate a multi-dimensional performance bottleneck diagnosis result of the target service system according to the interaction effect value of each combined dimension value.

[0130] The technical solution of an embodiment of the present invention collects call chain data in a target service and splits each call chain data into processing unit data, identifies target unit data that meets high time-consuming conditions in the processing unit data, wherein the target service includes multiple service nodes set in a set area, and these nodes are used to implement processing functions that match the set processing units in the call chain. Subsequently, a dimension list containing multiple dimensions is constructed according to the attributes of the target service, and each dimension contains multiple dimension values. The dimension importance of each dimension is calculated based on the processing time value of the target unit data under each dimension value, and a set number of target dimensions are screened out accordingly. Then, the dimension values ​​under at least two target dimensions are combined to obtain multiple combined dimension values, and the interaction effect value of each combined dimension value is calculated based on the processing time value of the target unit data under each combined dimension value. Finally, a multi-dimensional performance bottleneck diagnosis result is generated based on these interaction effect values. This new performance bottleneck diagnosis method effectively solves the problems of dimensional explosion and rough analysis faced by traditional methods in complex scenarios by establishing a systematic analysis process, providing a new technical path for performance optimization of distributed systems, significantly improving diagnosis efficiency while ensuring analysis accuracy.

[0131] Based on the above embodiments, a filtering module may be further included, wherein:

[0132] A filtering module is used to obtain the processing time value corresponding to each processing unit data before identification, and filter the processing unit data whose processing time value is outside a preset reasonable time range;

[0133] Based on the above embodiments, the data acquisition module 410 is specifically configured to:

[0134] Calculate the percentile of the target data indicator based on the processing time of each processing unit data;

[0135] The processing unit data whose processing time value exceeds the percentile under the target data indicator is extracted as the target unit data that meets the high time consumption condition.

[0136] Based on the above embodiments, the dimension screening module 430 may further include:

[0137] The baseline calculation submodule is used to obtain the current dimension in the dimension list and obtain the proportion of the target unit data in the processing unit data under the current dimension as the baseline high time consumption rate;

[0138] The segment high time consumption rate statistics submodule is used to obtain the proportion of target unit data in each processing unit data of each current dimension value under the current dimension as the segment high time consumption rate of each current dimension value;

[0139] The importance evaluation submodule is used to calculate the dimension importance of the current dimension based on the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each current dimension value under the current dimension.

[0140] Based on the above embodiments, the importance evaluation submodule is specifically used to:

[0141] Calculate the difference between the segment high time consumption rate and the baseline high time consumption rate of each current dimension value and multiply it by the number of processing unit data belonging to each current dimension value to obtain the information gain contribution value of each current dimension value;

[0142] The information gain contribution values ​​of each current dimension value are accumulated and summed to obtain the dimension importance of the current dimension.

[0143] Based on the above embodiments, the combined analysis module 440 is specifically configured to:

[0144] Combining the target dimensions according to a preset number of optional combinations to obtain at least one dimension set, wherein each dimension set contains at least two target dimensions;

[0145] Combining the dimensional values ​​of each target dimension in each dimensional set with the dimensional values ​​of other target dimensions in the dimensional set to obtain multiple combined candidate dimensional values;

[0146] The quantity value of each processing unit data belonging to each combined candidate dimension value is calculated respectively, and the combined candidate dimension value whose quantity value is greater than or equal to the preset quantity threshold is determined as the combined dimension value.

[0147] Based on the above embodiments, the combined analysis module 440 is specifically configured to:

[0148] Obtain the quantity value of the target unit data and the proportion of the target unit data in each processing unit data belonging to each combined dimension value;

[0149] Calculate the influence ratio corresponding to each combined dimension value based on the number of target unit data corresponding to each combined dimension value and the proportion of the target unit data;

[0150] In descending order of influence ratio, the influence ratios of each combined dimension value are accumulated and summed up. When the accumulated sum is greater than or equal to the target ratio value, the combined dimension values ​​currently participating in the calculation are used as the target combined dimension value.

[0151] Calculate the interaction effect value of each target combination dimension value based on the proportion of target unit data corresponding to each target combination dimension value and the expected proportion of target unit data;

[0152] The expected proportion of target unit data corresponding to the target combination dimension value is calculated by calculating the baseline high time consumption rate of each dimension in the target combination dimension value.

[0153] Based on the above embodiments, the diagnosis decision module 450 is specifically configured to:

[0154] identifying at least one abnormal pattern in each target combination dimension value based on the interaction effect value of each target combination dimension value;

[0155] Based on the identified abnormal patterns, multi-dimensional performance bottleneck diagnosis results of the target service system are generated.

[0156] The multi-dimensional performance bottleneck diagnosis device based on call chain provided by an embodiment of the present invention can execute the multi-dimensional performance bottleneck diagnosis method based on call chain provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0157] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0158] Example 5

[0159] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0160] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0161] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0162] The processor 11 may be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a multi-dimensional performance bottleneck diagnosis method based on a call chain, namely:

[0163] Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data;

[0164] The target service system includes multiple service nodes, which are set in a set area and are used to implement processing functions that match the set processing units in the call chain;

[0165] Constructing a dimension list of the target service system according to the system properties of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values;

[0166] Calculate the dimension importance of each dimension based on the processing time of the target unit data of each dimension value under each dimension, and filter out a set number of target dimensions based on the dimension importance;

[0167] Combining the dimensional values ​​of at least two target dimensions to obtain multiple combined dimensional values, and calculating the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value;

[0168] According to the interaction effect value of each combined dimension value, the multi-dimensional performance bottleneck diagnosis result of the target service system is generated.

[0169] In some embodiments, a multi-dimensional performance bottleneck diagnosis method based on a call chain can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the multi-dimensional performance bottleneck diagnosis method based on a call chain described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute a multi-dimensional performance bottleneck diagnosis method based on a call chain in any other appropriate manner (for example, by means of firmware).

[0170] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system comprising at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0171] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0172] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0173] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0174] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0175] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0176] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0177] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A multi-dimensional performance bottleneck diagnosis method based on call chain, characterized by: include: Collect call chain data in the target service system, split each collected call chain data into processing unit data, and identify target unit data that meets the high time consumption condition in each processing unit data; The target service system includes multiple service nodes, which are set in a set area and are used to implement processing functions that match the set processing units in the call chain; Constructing a dimension list of the target service system according to the system properties of the target service system, wherein the dimension list includes multiple dimensions, and each dimension includes multiple dimension values; Calculate the dimension importance of each dimension based on the processing time of the target unit data of each dimension value under each dimension, and filter out a set number of target dimensions based on the dimension importance; Combining the dimensional values ​​of at least two target dimensions to obtain multiple combined dimensional values, and calculating the interaction effect value of each combined dimensional value based on the processing time value of the target unit data belonging to each combined dimensional value; According to the interaction effect value of each combined dimension value, the multi-dimensional performance bottleneck diagnosis result of the target service system is generated.

2. The method according to claim 1, characterized in that Before identifying target unit data that meets the high time consumption condition in each processing unit data, the method further includes: Obtaining the processing time value corresponding to each processing unit data, and filtering the processing unit data whose processing time value is outside a preset reasonable time range; Identify target unit data that meets the high time consumption condition in each processing unit data, including: Calculate the percentile of the target data indicator based on the processing time of each processing unit data; The processing unit data whose processing time value exceeds the percentile under the target data indicator is extracted as the target unit data that meets the high time consumption condition.

3. The method according to claim 1, wherein: The dimension importance of each dimension is calculated based on the processing time of the target unit data of each dimension value under each dimension, including: Get the current dimension in the dimension list, and obtain the proportion of target unit data in each processing unit data under the current dimension as the baseline high time consumption rate; Obtain the proportion of target unit data in each processing unit data of each current dimension value under the current dimension as the segment height time consumption rate of each current dimension value; The dimension importance of the current dimension is calculated based on the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each current dimension value under the current dimension.

4. The method according to claim 3, characterized in that Calculate the dimension importance of the current dimension based on the baseline high time consumption rate of the current dimension and the segment high time consumption rate of each current dimension value under the current dimension, including: Calculate the difference between the segment high time consumption rate and the baseline high time consumption rate of each current dimension value and multiply it by the number of processing unit data belonging to each current dimension value to obtain the information gain contribution value of each current dimension value; The information gain contribution values ​​of each current dimension value are accumulated and summed to obtain the dimension importance of the current dimension.

5. The method according to claim 1, wherein Combine the dimension values ​​under at least two target dimensions to obtain multiple combined dimension values, including: Combining the target dimensions according to a preset number of optional combinations to obtain at least one dimension set, wherein each dimension set contains at least two target dimensions; Combining the dimensional values ​​of each target dimension in each dimensional set with the dimensional values ​​of other target dimensions in the dimensional set to obtain multiple combined candidate dimensional values; The quantity value of each processing unit data belonging to each combined candidate dimension value is calculated respectively, and the combined candidate dimension value whose quantity value is greater than or equal to the preset quantity threshold is determined as the combined dimension value.

6. The method according to any one of claims 1 to 5, characterized in that According to the processing time value of the target unit data belonging to each combined dimension value, the interaction effect value of each combined dimension value is calculated, including: Obtain the quantity value of the target unit data and the proportion of the target unit data in each processing unit data belonging to each combined dimension value; Calculate the influence ratio corresponding to each combined dimension value based on the number of target unit data corresponding to each combined dimension value and the proportion of the target unit data; In descending order of influence ratio, the influence ratios of each combined dimension value are accumulated and summed up. When the accumulated sum is greater than or equal to the target ratio value, the combined dimension values ​​currently participating in the calculation are used as the target combined dimension value. Calculate the interaction effect value of each target combination dimension value based on the proportion of target unit data corresponding to each target combination dimension value and the expected proportion of target unit data; The expected proportion of target unit data corresponding to the target combination dimension value is calculated by calculating the baseline high time consumption rate of each dimension in the target combination dimension value.

7. The method according to claim 6, characterized in that Based on the interaction effect values ​​of each combined dimension value, a multi-dimensional performance bottleneck diagnosis result of the target service system is generated, including: identifying at least one abnormal pattern in each target combination dimension value based on the interaction effect value of each target combination dimension value; Based on the identified abnormal patterns, multi-dimensional performance bottleneck diagnosis results of the target service system are generated.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the multi-dimensional performance bottleneck diagnosis method based on call chains according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the multi-dimensional performance bottleneck diagnosis method based on call chains according to any one of claims 1 to 7 when executed.

10. A computer program product, characterized in that The computer program product includes a computer program, which, when executed by a processor, implements the call chain-based multi-dimensional performance bottleneck diagnosis method according to any one of claims 1 to 7.

Images