Automatic analysis method and device of job chain and computer program product

Abstract

The invention discloses an automatic analysis method and device for a job chain and a computer program product. The method relates to the field of artificial intelligence, and comprises the following steps: determining a target operation chain to be analyzed, and obtaining topological structure information of the target operation chain and node table data of each target operation node in the target operation chain; the topological structure information and the node table data of each target job node are input into a target model, the target predicted operation duration of each target job node is output through the target model, and the target model comprises a first model, a second model and a third model; constructing a directed acyclic graph according to the target predicted operation duration of each target job node; identifying abnormal job nodes in the target job chain according to the directed acyclic graph; and generating an analysis report of the target job chain according to the directed acyclic graph and the abnormal job nodes. Through the method and the device, the problems of low efficiency and low accuracy of a traditional operation chain performance analysis method in related technologies are solved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and more specifically, to an automated analysis method, apparatus, and computer program product for a work chain. Background Technology

[0002] In today's data-intensive industries, such as finance, telecommunications, and large e-commerce, core business systems face the daily task of handling massive amounts of backend operations. These operations are often organized into tightly linked chains, with each link undertaking key functions from data extraction, cleaning, and transformation to data loading and business report generation. With the surge in data volume and the increase in business complexity, the scale of these operation chains grows daily, presenting a complex network of hundreds or even thousands of operation nodes. This leads to a significant increase in the overall runtime of the operation chain, sometimes even exceeding the predetermined "batch processing window."

[0003] To address these issues, IT operations and development teams typically employ traditional manual analysis methods, attempting to trace the root causes of performance bottlenecks retrospectively. However, this traditional approach, reliant on personal experience, lacks standardized processes. The conclusions and focus of each analysis can vary from person to person, leading to unreliability and non-reproducibility of the results. Furthermore, collecting and organizing workflow data is itself a time-consuming and error-prone process, requiring significant effort from operations personnel. Even when data is ready, manual analysis often focuses on the longest-running jobs, neglecting those jobs that, while having short individual run times, abnormally consume resources or exceed average execution times. These "abnormal execution data" are crucial to the overall efficiency of the workflow. In conclusion, traditional workflow performance analysis methods are not only inefficient but also suffer from significant deficiencies in analytical depth and scientific basis, failing to meet the increasingly sophisticated demands of modern enterprises for workflow performance management.

[0004] There is currently no effective solution to the problem that traditional job chain performance analysis methods in related technologies are inefficient and have low accuracy. Summary of the Invention

[0005] The main objective of this application is to provide an automated analysis method, apparatus, and computer program product for work chains, in order to solve the problems of low efficiency and low accuracy of traditional work chain performance analysis methods in related technologies.

[0006] To achieve the above objectives, according to one aspect of this application, an automated analysis method for job chains is provided. The method includes: determining the target job chain to be analyzed, and obtaining the topology information of the target job chain and node table data for each target job node in the target job chain; inputting the topology information and node table data for each target job node into a target model, and outputting the target predicted runtime for each target job node through the target model, wherein the target model includes a first model, a second model, and a third model, and the third model includes multiple sub-models for job nodes handling different business types; constructing a directed acyclic graph (DAG) based on the target predicted runtime for each target job node; identifying abnormal job nodes in the target job chain based on the DAG; and generating an analysis report of the target job chain based on the DAG and the abnormal job nodes.

[0007] Optionally, inputting the topology information and the node table data of each target job node into the target model, and outputting the target predicted runtime of each target job node through the target model includes: determining the first weight of the first model, the second weight of the second model, and the third weight of the third model; inputting the topology information and the node table data of each target job node into the first model, and outputting the first predicted runtime of each target job node through the first model; performing differential processing on the node table data of each target job node to obtain differentially processed table data; inputting the topology information and the differentially processed table data of each target job node into the second model, and outputting the second predicted runtime of each target job node through the second model; determining the business type of each target job node, inputting the business type of each job node into the target sub-model in the third model, and outputting the third predicted runtime of each target job node through the target sub-model; and weightedly fusing the first predicted runtime, the second predicted runtime, and the third predicted runtime according to the first weight, the second weight, and the third weight to obtain the target predicted runtime of each target job node.

[0008] Optionally, the target model is trained through the following steps: training a first initial model based on first training data to obtain a first model, wherein the first training data includes the topology information of the sample job chain, the node table data of each sample job node in the sample job chain, and the runtime of each sample job node; training a second initial model based on second training data to obtain a second model, wherein the second training data includes the topology information of the sample job chain, the differentially processed table data of each sample job node in the sample job chain, and the runtime of each sample job node; training a third initial model based on third training data to obtain a third model, wherein the third training data includes the topology information of sample job chains of multiple different business types, the node table data of each sample job node in the sample job chains of multiple different business types, and the runtime of each sample job node in the sample job chains of multiple different business types; and obtaining the target model based on the first model, the second model, and the third model.

[0009] Optionally, constructing a directed acyclic graph based on the target predicted runtime of each target job node includes: treating each target job node as a graph node in the directed acyclic graph, and treating the target predicted runtime of each target job node as the weight of each graph node; and constructing a directed acyclic graph based on each graph node and the weight of each graph node.

[0010] Optionally, identifying abnormal job nodes in the target job chain based on the directed acyclic graph includes: calculating the total time consumption of each path in the directed acyclic graph, identifying the target path in the directed acyclic graph, wherein the target path is the path with the longest total time consumption among the top k sorted paths; calculating the time difference between the target predicted runtime and the actual runtime of each target job node in the target path; calculating the resource utilization rate difference between the resource utilization rate and the historical resource utilization rate of each target job node in the target path; and identifying target job nodes whose time difference is greater than a first preset difference threshold and / or whose resource utilization rate difference is greater than a second preset difference threshold as abnormal job nodes.

[0011] Optionally, generating an analysis report for the target job chain based on the directed acyclic graph and abnormal job nodes includes: generating optimization suggestions through a rule engine based on the target path and abnormal job nodes; and generating an analysis report for the target job chain using a predefined analysis report template based on the target predicted runtime, target path, abnormal job nodes, and optimization suggestions for each target job node.

[0012] Optionally, after generating an analysis report of the target work chain based on the abnormal work nodes, the method further includes: optimizing the target work chain based on the analysis report and monitoring the optimization results of the optimized target work chain; and adjusting the target model based on the optimization results.

[0013] To achieve the above objectives, according to another aspect of this application, an automated analysis apparatus for job chains is provided. The apparatus includes: a first determining unit, configured to determine the target job chain to be analyzed and acquire topology information of the target job chain and node table data of each target job node in the target job chain; an input unit, configured to input the topology information and node table data of each target job node into a target model, and output the target predicted runtime of each target job node through the target model, wherein the target model includes a first model, a second model, and a third model, and the third model includes multiple sub-models of job nodes processing different business types; a construction unit, configured to construct a directed acyclic graph (DAG) based on the target predicted runtime of each target job node; an identification unit, configured to identify abnormal job nodes in the target job chain based on the DAG; and a generation unit, configured to generate an analysis report of the target job chain based on the DAG and the abnormal job nodes.

[0014] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include methods for performing any one of the methods.

[0015] According to another aspect of this application, a computer program product is provided, including computer instructions that, when executed by a processor, implement the steps of any of the above methods.

[0016] In this embodiment, the target job chain to be analyzed is determined, and the topology information of the target job chain and the node table data of each target job node in the target job chain are obtained. The topology information and the node table data of each target job node are input into the target model, and the target predicted runtime of each target job node is output through the target model. The target model includes a first model, a second model, and a third model. The third model includes multiple sub-models of job nodes that process different business types. A directed acyclic graph is constructed based on the target predicted runtime of each target job node. Abnormal job nodes in the target job chain are identified based on the directed acyclic graph. An analysis report of the target job chain is generated based on the directed acyclic graph and the abnormal job nodes. This solves the technical problems of low efficiency and low accuracy of traditional job chain performance analysis methods. In this application, the topology of the target job chain and the node table data of each target job node are input into the trained target model. The first, second and third models in the target model output the predicted runtime of each target job node, respectively. The predicted runtimes obtained by fusing the three models are used to obtain the target predicted runtime of each target job node. A directed acyclic graph is constructed based on the target predicted runtime of each target job node. By calculating and identifying the target path in the graph and identifying abnormal job nodes on the target path, an analysis report of the target job chain is finally generated, which achieves the technical effect of improving the efficiency and accuracy of automated job chain analysis. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A hardware block diagram of a computer terminal for implementing an automated analysis method for job chains is shown.

[0019] Figure 2 This is a flowchart of an automated analysis method for a work chain according to an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of an automated analysis device for a work chain according to an embodiment of this application;

[0021] Figure 4 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] Example 1

[0025] According to an embodiment of this application, an embodiment of a method for automated analysis of a job chain is also provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal (or mobile device) for implementing an automated analysis method for job chains is shown. Figure 1As shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0027] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0028] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the automated analysis method of the job chain in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned automated analysis method of the job chain. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0029] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0030] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0031] Under the aforementioned operating environment, this application provides the following: Figure 2 The method for automating the analysis of the work chain is shown. Figure 2 This is a flowchart of an automated analysis method for a work chain according to Embodiment 1 of this application.

[0032] Step S201: Determine the target job chain to be analyzed, and obtain the topology information of the target job chain and the node table data of each target job node in the target job chain.

[0033] Optionally, the topology information of the target job chain and the node table data of each target job node can be obtained from a dedicated time-series database. First, the workflow connects to the job scheduling system and IT monitoring system through a pre-defined API interface and database connector. Then, the workflow can periodically (e.g., daily) extract static and dynamic time-series data from the data source. The static data includes at least the topology and dependencies of the job chain, while the dynamic time-series data includes at least the historical runtime of each job node, central memory / memory usage, I / O throughput, and success / failure status records. Finally, the extracted raw data is standardized and stored in a dedicated time-series database to form the dataset required for subsequent analysis. The topology information of the target job chain is a graphical representation of the dependencies between each target job node in the target job chain. It can be a directed graph, where the arrows indicate the direction of data flow and the order of job execution. The node table data of each target job node includes the ID, name, business type, running frequency, historical runtime, resource consumption data, and other relevant indicators of each target job node.

[0034] Step S202: Input the topology information and the node table data of each target job node into the target model, and output the target predicted runtime of each target job node through the target model. The target model includes a first model, a second model and a third model. The third model includes multiple sub-models of job nodes that process different business types.

[0035] Optionally, the topology information and node table data collected in step S201 are used as input to a pre-trained target model. The target model consists of three main parts: a first model (for raw time-series data), a second model (for differential time-series data), and a third model (for data in specific business scenarios). Then, the target model integrates the outputs of the above three models to obtain the target predicted runtime of each target job node in the target job chain, that is, the estimated time required for the next run.

[0036] Step S203: Construct a directed acyclic graph based on the target predicted runtime of each target job node.

[0037] Optionally, a directed acyclic graph (DAG) can be constructed using the target predicted runtime of each target job node and the topological information of the target job chain. In the DAG, each node represents a target job node, and edges represent the dependencies between target job nodes. The weight of a node is its target predicted runtime, reflecting the estimated time consumed by the target job node in its next run.

[0038] Step S204: Identify abnormal job nodes in the target job chain based on the directed acyclic graph.

[0039] Optionally, based on the constructed directed acyclic graph, abnormal job nodes can be identified. The predicted target runtime of each target job node can be compared with the acquired historical runtime to find job nodes whose time consumption deviation exceeds a preset threshold, as well as job nodes with significant deviations in resource consumption, i.e., abnormal job nodes.

[0040] Step S205: Generate an analysis report of the target job chain based on the directed acyclic graph and abnormal job nodes.

[0041] Optionally, after identifying abnormal job nodes and the critical path in the directed acyclic graph, the report generation stage can begin. By summarizing and structuring the analysis results, including details of the critical path, specific information about abnormal nodes, data points for deviation analysis, and corresponding optimization suggestions, an analysis report can be automatically generated. This report can be distributed to the IT operations and development teams via email or instant messaging tools to ensure they receive detailed analysis results promptly.

[0042] The automated job chain analysis method provided in this application identifies the target job chain to be analyzed and obtains its topology information and node table data for each target job node. The topology information and node table data are input into a target model, which outputs the target predicted runtime for each target job node. The target model includes a first model, a second model, and a third model, with the third model containing multiple sub-models for job nodes handling different business types. A directed acyclic graph (DAG) is constructed based on the target predicted runtime for each target job node. Abnormal job nodes in the target job chain are identified based on the DAG. An analysis report of the target job chain is generated based on the DAG and the abnormal job nodes. This method solves the technical problems of low efficiency and low accuracy in traditional job chain performance analysis methods. In this application, the topology of the target job chain and the node table data of each target job node are input into the trained target model. The first, second and third models in the target model output the predicted runtime of each target job node, respectively. The predicted runtimes obtained by fusing the three models are used to obtain the target predicted runtime of each target job node. A directed acyclic graph is constructed based on the target predicted runtime of each target job node. By calculating and identifying the target path in the graph and identifying abnormal job nodes on the target path, an analysis report of the target job chain is finally generated, which achieves the technical effect of improving the efficiency and accuracy of automated job chain analysis.

[0043] To improve the accuracy of the predicted runtime of each target job node in the target model, optionally, in the automated job chain analysis method provided in this application embodiment, the topology information and the node table data of each target job node are input into the target model, and the target predicted runtime of each target job node is output by the target model, including:

[0044] The first step is to determine the first weight of the first model, the second weight of the second model, and the third weight of the third model.

[0045] Optionally, the first model processes the raw time-series data, the second model processes the differential time-series data, and the third model includes multiple sub-models that process time-series data for specific business scenarios. Each model contributes differently to the final prediction result due to differences in the data type, business scenario, or algorithm characteristics it processes. Determining the weight of each model reflects the relative importance of its prediction in the final integrated prediction. A higher weight indicates a larger proportion of the model's prediction in the final fusion result and a stronger influence on the decision.

[0046] The second step is to input the topology information and the node table data of each target job node into the first model, and then output the first predicted runtime of each target job node through the first model.

[0047] Optionally, the first model is trained based on the overall historical data of the job chain and can make a preliminary prediction of the runtime of each job node. The data input to the first model includes the topology information of the target job chain and the original node table data. The first model can analyze this data through its internal machine learning algorithm to predict the runtime of each target job node and obtain the first predicted runtime.

[0048] The third step is to perform differential processing on the node table data of each target task node to obtain the differentially processed table data.

[0049] Optionally, differential processing is a data preprocessing technique that calculates the difference between adjacent data points, allowing attention to be focused on the rate of change rather than the absolute value of the data. Differential processing of the node table data for each target job node yields a differentially processed table data that better reflects the volatility and trend of the runtime of each target job node. This helps the second model capture the instantaneous changes in the runtime of each target job node, improving its predictive ability for short-term emergencies.

[0050] The fourth step involves inputting the topology information and the differentially processed table data of each target job node into the second model, and then outputting the second predicted runtime of each target job node through the second model.

[0051] Optionally, the second model is used to analyze the differential data of job node runtime. It receives the differentially processed node table data as input and uses its own algorithm to predict the runtime of each target job node. The prediction results of the second model reflect the short-term trend and are helpful in identifying sudden performance degradation or surges in resource demand.

[0052] The fifth step is to determine the business type of each target job node, input the target sub-model in the third model according to the business type of each job node, and output the third predicted runtime of each target job node through the target sub-model.

[0053] Optionally, since the business types and data characteristics processed by different target job nodes in the target job chain may differ, the third model includes multiple sub-models, each trained for a target job node with a specific business type. A target sub-model can be selected from the third model based on the business type of each target job node, and then used to predict the runtime of the target job node, thus obtaining the third predicted runtime.

[0054] The sixth step is to weight and fuse the first, second, and third predicted runtimes based on the first, second, and third weights to obtain the target predicted runtime for each target job node.

[0055] Optionally, the prediction results of the three models (first predicted runtime, second predicted runtime, and third predicted runtime) are weighted and fused together. The three prediction results are combined using a weighted summation formula to obtain the final target predicted runtime for each target job node.

[0056] In summary, through the above steps, the target model predicts the runtime of each target job node from three aspects: the macro-characteristics of the target job chain, the short-term trend of change, and the specific business type. The precise prediction of the runtime of each target job node in the target job chain can improve the accuracy of the performance analysis of the target job chain.

[0057] To improve the accuracy of the target model in predicting the runtime of job nodes, optionally, in the automated job chain analysis method provided in this application embodiment, the target model is trained through the following steps:

[0058] The first step is to train a first initial model based on the first training data to obtain the first model. The first training data includes the topological structure information of the sample job chain, the node table data of each sample job node in the sample job chain, and the runtime of each sample job node.

[0059] Optionally, the first training data includes the topology information of the sample job chains, the node table data of each sample job node, and the actual runtime of these nodes. During training, firstly, the topology information of the sample job chains and the node table data of each sample job node are input into the first initial model. Based on the current parameter settings, the first initial model predicts the runtime of each sample job node, generating preliminary prediction results. Next, the difference between the prediction results and the actual runtime is calculated, i.e., the prediction error. Then, the error is fed back into the model through a backpropagation algorithm to adjust the model parameters in reverse, thereby reducing the prediction error. This process is repeated multiple times, with each iteration aiming to reduce the prediction error and optimize the model parameters. The training process is set with a stopping condition, such as reaching a predetermined number of iterations or the prediction error being less than a certain threshold. When the stopping condition is met, the first model is obtained.

[0060] The second step is to train a second initial model based on the second training data to obtain a second model. The second training data includes the topological structure information of the sample job chain, the differentially processed table data of each sample job node in the sample job chain, and the runtime of each sample job node.

[0061] Optionally, the training process for the second model is similar to that of the first model, except that the training data is different. In the second training data, the node table data for each sample job node has been differentially processed. After training, a second model that can focus on short-term fluctuations is obtained.

[0062] The third step is to train the third initial model based on the third training data to obtain the third model. The third training data includes the topology information of sample job chains of multiple different business types, the node table data of each sample job node in the sample job chains of multiple different business types, and the runtime of each sample job node in the sample job chains of multiple different business types.

[0063] Optionally, the training process for the third model is similar to that of the first and second models, except that the training data differs. The third training data includes job chain data from multiple different business types. Each sub-model in the third model learns the operational rules and resource consumption patterns of a specific domain based on the different business types. After training, multiple sub-models capable of handling specific business scenarios are obtained.

[0064] The fourth step is to obtain the target model based on the first model, the second model, and the third model.

[0065] Optionally, after obtaining the first model, the second model, and the third model, the three models can be integrated into a target model through a model fusion mechanism, including but not limited to weighted average, voting mechanism, or more complex neural network fusion mechanisms.

[0066] In summary, by training the first, second, and third models through the above steps, and then fusing them to obtain the target model, we can leverage the strengths of each model to obtain a more comprehensive and powerful target model with stronger overall predictive capabilities.

[0067] To accurately analyze the performance of a target job chain based on the runtime of each target job node, optionally, in the automated job chain analysis method provided in this application embodiment, constructing a directed acyclic graph based on the target predicted runtime of each target job node includes:

[0068] The first step is to treat each target job node as a graph node in a directed acyclic graph, and to use the target predicted runtime of each target job node as the weight of each graph node.

[0069] Alternatively, a directed acyclic graph (DAG) is a graph structure where nodes are connected by directed edges, which can be used to visualize dependencies and runtime distributions in a target job chain. In a DAG, each graph node has a numerical attribute, which in this case refers to the target predicted runtime of the target job node.

[0070] The second step is to construct a directed acyclic graph based on each graph node and its weight.

[0071] Optionally, after obtaining each graph node (target job node) and its corresponding weight (target prediction runtime), directed edges are drawn based on the dependencies between target job nodes to connect the relevant nodes.

[0072] In summary, through the above steps, a directed acyclic graph is constructed, which presents the structure of the target job chain. The target predicted runtime of each target job node is also embedded in the node weight, providing a graphical basis for subsequent performance analysis.

[0073] To more comprehensively assess the health status of the target job chain, optionally, in the automated job chain analysis method provided in this application embodiment, identifying abnormal job nodes in the target job chain based on a directed acyclic graph includes:

[0074] The first step is to calculate the total time of each path in the directed acyclic graph and identify the target path in the directed acyclic graph, where the target path is the path with the longest total time among the first k sorted paths.

[0075] Optionally, by traversing all possible paths in the directed acyclic graph and calculating the sum of the node weights on each path, the total time taken for that path can be obtained. Then, all paths are sorted according to the calculated total time, and the top k longest paths are selected as the target path.

[0076] The second step is to calculate the difference between the predicted runtime and the actual runtime of each target job node in the target path.

[0077] Optionally, the duration difference refers to the gap between the target predicted runtime and the actual runtime, which can be used to assess the deviation of the job node's running status. For each job node identified as being on the target path, its target predicted runtime (predicted by the target model) is compared with its actual runtime (obtained from the monitoring system), and the difference between the two is calculated, which can quantify the degree of running deviation of each job node relative to the expected.

[0078] The third step is to calculate the difference between the resource utilization rate and the historical resource utilization rate of each target job node in the target path.

[0079] Optionally, the resource utilization difference refers to the deviation between the current resource utilization rate and the historical resource utilization rate (which can be the average or median), used to determine whether resource consumption is abnormal. By comparing the resource utilization rate in the current operating cycle with the historical average resource utilization rate, job nodes with a sudden increase in resource consumption can be identified.

[0080] The fourth step is to identify target job nodes whose duration difference is greater than the first preset difference threshold and / or whose resource utilization difference is greater than the second preset difference threshold as abnormal job nodes.

[0081] Optionally, the first preset difference threshold is the boundary used to measure whether the duration difference reaches the abnormal standard, and the second preset difference threshold is the boundary used to determine whether the resource utilization difference exceeds the normal fluctuation range. Combining the duration difference and resource utilization difference calculated in the first two steps, two thresholds (the first preset difference threshold and the second preset difference threshold) are set. Target job nodes whose duration difference or resource utilization difference exceeds the corresponding threshold are marked as abnormal job nodes, ensuring that only those job nodes that truly deviate from the normal operating state are marked, avoiding overreaction to minor fluctuations.

[0082] In summary, by following the steps above, the most time-consuming path in the target job chain can be identified, and the target job node with unexpectedly long processing time or abnormal resource consumption can be located.

[0083] To efficiently generate automated analysis reports for target job chains, optionally, in the automated job chain analysis method provided in this application embodiment, generating analysis reports for target job chains based on directed acyclic graphs and abnormal job nodes includes:

[0084] The first step is to generate optimization suggestions based on the target path and abnormal job nodes using the rule engine;

[0085] Optionally, the rules engine can automatically judge and generate decision or optimization suggestions based on predefined rules or business logic, such as adjusting resource allocation, optimizing code logic, and improving job scheduling strategies.

[0086] The second step is to generate an analysis report for the target job chain based on the target predicted runtime, target path, abnormal job nodes, and optimization suggestions for each target job node, using a predefined analysis report template.

[0087] Optionally, the target path, abnormal job nodes, and optimization suggestions can be summarized and organized, and all structured data can be rendered into a predefined report template to generate an analysis report. The analysis report includes, but is not limited to: a report summary; target path analysis, including all job nodes on the target path, the predicted time of each job node, and the total time; an abnormal operation data list, which can list all identified abnormal job nodes in tabular form, and indicate the abnormality type (time deviation, resource deviation, etc.) and specific values; and an optimization suggestion list, which lists specific and actionable optimization suggestions one by one, and associates them with the corresponding critical path or abnormal data.

[0088] In summary, by following the steps outlined above and analyzing each target task node in the target workflow, analysis reports can be generated automatically and efficiently.

[0089] To achieve continuous optimization of the target job chain, optionally, in the automated job chain analysis method provided in this application embodiment, after generating an analysis report of the target job chain based on abnormal job nodes, the method further includes:

[0090] The first step is to optimize the target workflow based on the analysis report and monitor the optimization results of the target workflow.

[0091] Optionally, based on the optimization suggestions proposed in the automated analysis report, the IT operations or development team will conduct targeted optimizations to the target workflow. After implementing the optimization measures, the operational status of the target workflow will be continuously monitored, and the optimization results will be collected to evaluate the effectiveness of the optimization measures.

[0092] The second step is to adjust the target model based on the optimization results.

[0093] Optionally, the target model can be retrospectively adjusted based on the optimized target job chain operation data (i.e., the optimization results) to enhance its prediction accuracy and adaptability.

[0094] In summary, through the above steps, from generating the analysis report to optimizing the target work chain and adjusting the target model, a coherent improvement chain is formed, ensuring the continuous optimization of the target work chain performance and promoting the continuous improvement of the target model.

[0095] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0096] Example 2

[0097] This application also provides an automated analysis device for a work chain. It should be noted that this automated analysis device can be used to execute the automated analysis method for work chains provided in this application. The automated analysis device for a work chain provided in this application is described below.

[0098] According to embodiments of this application, an apparatus for implementing the above-described automated analysis method for the work chain is also provided, such as... Figure 3 As shown, the device includes: a first determining unit 301, an input unit 302, a constructing unit 303, an identification unit 304, and a generating unit 305.

[0099] Specifically, the first determining unit 301 is used to determine the target job chain to be analyzed, and to obtain the topological structure information of the target job chain and the node table data of each target job node in the target job chain.

[0100] The input unit 302 is used to input the topology information and the node table data of each target job node into the target model, and output the target predicted runtime of each target job node through the target model. The target model includes a first model, a second model and a third model. The third model includes multiple sub-models for job nodes that process different business types.

[0101] Construction unit 303 is used to construct a directed acyclic graph based on the target predicted runtime of each target job node;

[0102] The identification unit 304 is used to identify abnormal job nodes in the target job chain based on the directed acyclic graph.

[0103] The generation unit 305 is used to generate an analysis report of the target job chain based on the directed acyclic graph and abnormal job nodes.

[0104] The automated job chain analysis device provided in this application embodiment determines the target job chain to be analyzed through a first determining unit 301, and obtains the topology information of the target job chain and the node table data of each target job node in the target job chain; an input unit 302 inputs the topology information and the node table data of each target job node into a target model, and outputs the target predicted runtime of each target job node through the target model, wherein the target model includes a first model, a second model and a third model, and the third model includes multiple sub-models of job nodes that process different business types; a construction unit 303 constructs a directed acyclic graph based on the target predicted runtime of each target job node; an identification unit 304 identifies abnormal job nodes in the target job chain based on the directed acyclic graph; and a generation unit 305 generates an analysis report of the target job chain based on the directed acyclic graph and the abnormal job nodes. This solves the problems of low efficiency and low accuracy of traditional job chain performance analysis methods in related technologies, and achieves the effect of improving the efficiency and accuracy of automated job chain analysis.

[0105] Optionally, in the automated analysis device for the job chain provided in this application embodiment, the input unit 302 includes: a first determining module, used to determine the first weight of the first model, the second weight of the second model, and the third weight of the third model; a first input module, used to input the topology information and the node table data of each target job node into the first model, and output the first predicted runtime of each target job node through the first model; a differential processing module, used to perform differential processing on the node table data of each target job node to obtain differentially processed table data; a second input module, used to input the topology information and the differentially processed table data of each target job node into the second model, and output the second predicted runtime of each target job node through the second model; a third input module, used to determine the business type of each target job node, input the target sub-model in the third model according to the business type of each job node, and output the third predicted runtime of each target job node through the target sub-model; and a second determining module, used to weightedly fuse the first predicted runtime, the second predicted runtime, and the third predicted runtime according to the first weight, the second weight, and the third weight to obtain the target predicted runtime of each target job node.

[0106] Optionally, in the automated analysis device for job chains provided in this application embodiment, the device further includes: a first training unit, configured to train a first initial model based on first training data to obtain a first model, wherein the first training data includes topology information of sample job chains, node table data of each sample job node in the sample job chain, and runtime of each sample job node; a second training unit, configured to train a second initial model based on second training data to obtain a second model, wherein the second training data includes topology information of sample job chains, differentially processed table data of each sample job node in the sample job chain, and runtime of each sample job node; a third training unit, configured to train a third initial model based on third training data to obtain a third model, wherein the third training data includes topology information of multiple sample job chains of different business types, node table data of each sample job node in multiple sample job chains of different business types, and runtime of each sample job node in multiple sample job chains of different business types; and a second determining unit, configured to obtain a target model based on the first model, the second model, and the third model.

[0107] Optionally, in the automated analysis device for the job chain provided in the embodiments of this application, the construction unit includes: a third determining module, used to treat each target job node as a graph node in a directed acyclic graph, and to treat the target predicted runtime of each target job node as the weight of each graph node; and a construction module, used to construct a directed acyclic graph based on each graph node and the weight of each graph node.

[0108] Optionally, in the automated analysis device for the job chain provided in this application embodiment, the identification unit 304 includes: a first calculation module, used to calculate the total time consumption of each path in the directed acyclic graph and identify the target path in the directed acyclic graph, wherein the target path is the path with the longest total time consumption among the top k sorted paths; a second calculation module, used to calculate the time difference between the target predicted runtime and the actual runtime of each target job node in the target path; a third calculation module, used to calculate the resource utilization rate difference between the resource utilization rate and the historical resource utilization rate of each target job node in the target path; and a fourth determination module, used to determine the target job nodes with a time difference greater than a first preset difference threshold and / or the target job nodes with a resource utilization rate difference greater than a second preset difference threshold as abnormal job nodes.

[0109] Optionally, in the automated analysis device for the job chain provided in this application embodiment, the generation unit 305 includes: a first generation module, used to generate optimization suggestions through a rule engine based on the target path and abnormal job nodes; and a second generation module, used to generate an analysis report of the target job chain based on the target predicted runtime, target path, abnormal job nodes and optimization suggestions of each target job node, using a predefined analysis report template.

[0110] Optionally, in the automated analysis device for the work chain provided in the embodiments of this application, the device further includes: an optimization unit, used to optimize the target work chain according to the analysis report after generating an analysis report of the target work chain based on the abnormal work nodes, and to monitor the optimization results of the optimized target work chain; and an adjustment unit, used to adjust the target model according to the optimization results.

[0111] It should be noted that the first determining unit 301, input unit 302, construction unit 303, identification unit 304, and generation unit 305 mentioned above correspond to steps S201 to S205 in Embodiment 1. The five units and the corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules or units can be hardware or software components stored in memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n). The above modules can also be part of a device and run in the computer terminal 10 provided in Embodiment 1.

[0112] Example 3

[0113] Embodiments of this application may provide a computer terminal, which may be any computer terminal device in a group of computer terminals. Optionally, in this embodiment, the aforementioned computer terminal may also be replaced with a mobile terminal or an electronic device, etc.

[0114] Optionally, in this embodiment, the computer terminal may be located in at least one of a plurality of network devices in a computer network.

[0115] In this embodiment, the computer terminal described above can execute the program code for the following steps in the automated analysis method for job chains: determining the target job chain to be analyzed, and obtaining the topology information of the target job chain and the node table data of each target job node in the target job chain; inputting the topology information and the node table data of each target job node into the target model, and outputting the target predicted runtime of each target job node through the target model, wherein the target model includes a first model, a second model, and a third model, and the third model includes multiple sub-models of job nodes that process different business types; constructing a directed acyclic graph based on the target predicted runtime of each target job node; identifying abnormal job nodes in the target job chain based on the directed acyclic graph; and generating an analysis report of the target job chain based on the directed acyclic graph and the abnormal job nodes.

[0116] Optionally, the aforementioned computer terminal can execute program code for the following steps in the automated analysis method for the job chain: inputting topology information and node table data of each target job node into the target model, and outputting the target predicted runtime of each target job node through the target model, including: determining the first weight of the first model, the second weight of the second model, and the third weight of the third model; inputting topology information and node table data of each target job node into the first model, and outputting the first predicted runtime of each target job node through the first model; performing differential processing on the node table data of each target job node to obtain differentially processed table data; inputting topology information and differentially processed table data of each target job node into the second model, and outputting the second predicted runtime of each target job node through the second model; determining the business type of each target job node, inputting the business type of each job node into the target sub-model in the third model, and outputting the third predicted runtime of each target job node through the target sub-model; and weightedly fusing the first predicted runtime, the second predicted runtime, and the third predicted runtime according to the first weight, the second weight, and the third weight to obtain the target predicted runtime of each target job node.

[0117] Optionally, the aforementioned computer terminal can execute program code for the following steps in the automated analysis method for job chains: The target model is trained through the following steps: a first initial model is trained based on first training data to obtain a first model, wherein the first training data includes the topology information of the sample job chains, the node table data of each sample job node in the sample job chains, and the runtime of each sample job node; a second initial model is trained based on second training data to obtain a second model, wherein the second training data includes the topology information of the sample job chains, the differentially processed table data of each sample job node in the sample job chains, and the runtime of each sample job node; a third initial model is trained based on third training data to obtain a third model, wherein the third training data includes the topology information of sample job chains of multiple different business types, the node table data of each sample job node in the sample job chains of multiple different business types, and the runtime of each sample job node in the sample job chains of multiple different business types; and the target model is obtained based on the first model, the second model, and the third model.

[0118] Optionally, the computer terminal described above can execute program code for the following steps in the automated analysis method of the job chain: constructing a directed acyclic graph based on the target predicted runtime of each target job node, including: treating each target job node as a graph node in the directed acyclic graph, and treating the target predicted runtime of each target job node as the weight of each graph node; constructing a directed acyclic graph based on each graph node and the weight of each graph node.

[0119] Optionally, the aforementioned computer terminal can execute program code for the following steps in the automated analysis method for job chains: identifying abnormal job nodes in the target job chain based on the directed acyclic graph includes: calculating the total time consumption of each path in the directed acyclic graph, identifying the target path in the directed acyclic graph, wherein the target path is the path with the longest total time consumption among the top k sorted paths; calculating the time difference between the target predicted runtime and the actual runtime of each target job node in the target path; calculating the resource utilization rate difference between the resource utilization rate and the historical resource utilization rate of each target job node in the target path; and identifying target job nodes whose time difference is greater than a first preset difference threshold and / or whose resource utilization rate difference is greater than a second preset difference threshold as abnormal job nodes.

[0120] Optionally, the computer terminal described above can execute program code for the following steps in the automated analysis method for job chains: generating an analysis report for the target job chain based on the directed acyclic graph and abnormal job nodes, including: generating optimization suggestions through a rule engine based on the target path and abnormal job nodes; and generating an analysis report for the target job chain using a predefined analysis report template based on the target predicted runtime, target path, abnormal job nodes, and optimization suggestions for each target job node.

[0121] Optionally, the computer terminal described above can execute program code for the following steps in the automated analysis method of the job chain: after generating an analysis report of the target job chain based on abnormal job nodes, the method further includes: optimizing the target job chain based on the analysis report and monitoring the optimization results of the optimized target job chain; and adjusting the target model based on the optimization results.

[0122] Optionally, Figure 4 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 4 As shown, the electronic device may include: one or more ( Figure 4 (Only one is shown) Processor 402, memory 404, memory controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0123] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the automated analysis method and apparatus for the job chain in this embodiment. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned automated analysis method for the job chain. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0124] The processor can access the information and application programs stored in the memory via the transmission device to execute the steps described above in the automated analysis method of the aforementioned work chain.

[0125] This application provides an automated analysis scheme for job chains. It involves identifying the target job chain to be analyzed and obtaining its topology information and node table data for each target job node. The topology information and node table data are input into a target model, which outputs the target predicted runtime for each target job node. The target model includes a first model, a second model, and a third model, with the third model containing multiple sub-models for job nodes handling different business types. A directed acyclic graph (DAG) is constructed based on the target predicted runtime for each target job node. Abnormal job nodes in the target job chain are identified based on the DAG. An analysis report of the target job chain is generated based on the DAG and the abnormal job nodes. This solution addresses the technical problems of low efficiency and low accuracy in traditional job chain performance analysis methods, achieving the technical effect of improving the efficiency and accuracy of automated job chain analysis.

[0126] Those skilled in the art will understand that Figure 4 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones, tablets, handheld computers, mobile internet devices (MIDs), PADs, and other terminal devices. Figure 4 This does not limit the structure of the aforementioned electronic device. For example, electronic devices may also include components that are more... Figure 4 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 4 The different configurations shown.

[0127] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0128] Example 4

[0129] Embodiments of this application also provide a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the automated analysis method for the job chain provided in Embodiment 1.

[0130] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0131] Embodiments of this application also provide a computer program product. Optionally, in this embodiment, the computer program product can be used to execute the program code of the automated analysis method for the job chain provided in Embodiment 1.

[0132] Optionally, in this embodiment, when the computer program product is executed on a data processing device, it is suitable to perform the following steps: determining the target job chain to be analyzed, and obtaining the topology information of the target job chain and the node table data of each target job node in the target job chain; inputting the topology information and the node table data of each target job node into the target model, and outputting the target predicted runtime of each target job node through the target model, wherein the target model includes a first model, a second model and a third model, and the third model includes multiple sub-models of job nodes that process different business types; constructing a directed acyclic graph based on the target predicted runtime of each target job node; identifying abnormal job nodes in the target job chain based on the directed acyclic graph; and generating an analysis report of the target job chain based on the directed acyclic graph and the abnormal job nodes.

[0133] Optionally, when the computer program product is executed on a data processing device, it is suitable to perform the following steps: inputting topology information and node table data of each target job node into the target model, and outputting the target predicted runtime of each target job node through the target model, including: determining the first weight of the first model, the second weight of the second model, and the third weight of the third model; inputting topology information and node table data of each target job node into the first model, and outputting the first predicted runtime of each target job node through the first model; performing differential processing on the node table data of each target job node to obtain differentially processed table data; inputting topology information and differentially processed table data of each target job node into the second model, and outputting the second predicted runtime of each target job node through the second model; determining the business type of each target job node, inputting the business type of each job node into the target sub-model in the third model, and outputting the third predicted runtime of each target job node through the target sub-model; and weightedly fusing the first predicted runtime, the second predicted runtime, and the third predicted runtime according to the first weight, the second weight, and the third weight to obtain the target predicted runtime of each target job node.

[0134] Optionally, when the computer program product is executed on a data processing device, it is suitable for executing the following program code: the target model is trained through the following steps: a first initial model is trained based on first training data to obtain a first model, wherein the first training data includes the topology information of the sample job chain, the node table data of each sample job node in the sample job chain, and the runtime of each sample job node; a second initial model is trained based on second training data to obtain a second model, wherein the second training data includes the topology information of the sample job chain, the differentially processed table data of each sample job node in the sample job chain, and the runtime of each sample job node; a third initial model is trained based on third training data to obtain a third model, wherein the third training data includes the topology information of sample job chains of multiple different business types, the node table data of each sample job node in the sample job chains of multiple different business types, and the runtime of each sample job node in the sample job chains of multiple different business types; and a target model is obtained based on the first model, the second model, and the third model.

[0135] Optionally, when the computer program product is executed on a data processing device, the program code suitable for performing the following steps: constructing a directed acyclic graph based on the target predicted runtime of each target job node includes: treating each target job node as a graph node in the directed acyclic graph, and treating the target predicted runtime of each target job node as the weight of each graph node; constructing a directed acyclic graph based on each graph node and the weight of each graph node.

[0136] Optionally, when the computer program product is executed on a data processing device, the program code suitable for performing the following steps: identifying abnormal job nodes in the target job chain based on the directed acyclic graph includes: calculating the total time duration of each path in the directed acyclic graph, identifying the target path in the directed acyclic graph, wherein the target path is the path with the longest total time duration among the top k sorted paths; calculating the time difference between the target predicted runtime and the actual runtime of each target job node in the target path; calculating the resource utilization rate difference between the resource utilization rate and the historical resource utilization rate of each target job node in the target path; and identifying target job nodes whose time difference is greater than a first preset difference threshold and / or whose resource utilization rate difference is greater than a second preset difference threshold as abnormal job nodes.

[0137] Optionally, when the computer program product is executed on a data processing device, the program code is suitable for performing the following steps: generating an analysis report of the target job chain based on the directed acyclic graph and abnormal job nodes, including: generating optimization suggestions through a rule engine based on the target path and abnormal job nodes; and generating an analysis report of the target job chain using a predefined analysis report template based on the target predicted runtime, target path, abnormal job nodes, and optimization suggestions for each target job node.

[0138] Optionally, when the computer program product is executed on a data processing device, the program code is suitable for performing the following steps: after generating an analysis report of the target job chain based on the abnormal job nodes, the method further includes: optimizing the target job chain based on the analysis report and monitoring the optimization results of the optimized target job chain; adjusting the target model based on the optimization results.

[0139] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0140] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0141] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0142] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0144] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0145] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An automated analysis method for a work chain, characterized in that, include: Identify the target job chain to be analyzed, and obtain the topology information of the target job chain and the node table data of each target job node in the target job chain; The topology information and the node table data of each target job node are input into the target model, and the target predicted runtime of each target job node is output through the target model. The target model includes a first model, a second model and a third model, and the third model includes multiple sub-models for job nodes that process different business types. Construct a directed acyclic graph based on the target predicted runtime of each target job node; Abnormal job nodes in the target job chain are identified based on the directed acyclic graph. An analysis report of the target job chain is generated based on the directed acyclic graph and the abnormal job nodes.

2. The method according to claim 1, characterized in that, The topology information and the node table data of each target job node are input into the target model, and the target predicted runtime of each target job node is output by the target model, including: Determine the first weight of the first model, the second weight of the second model, and the third weight of the third model; The topology information and the node table data of each target job node are input into the first model, and the first predicted runtime of each target job node is output through the first model. Differential processing is performed on the node table data of each target task node to obtain the differential table data; The topology information and the differentially processed table data of each target job node are input into the second model, and the second predicted runtime of each target job node is output through the second model. Determine the business type of each target job node, input the target sub-model in the third model according to the business type of each job node, and output the third predicted runtime of each target job node through the target sub-model; Based on the first weight, the second weight, and the third weight, the first predicted runtime, the second predicted runtime, and the third predicted runtime are weighted and fused to obtain the target predicted runtime for each target job node.

3. The method according to claim 1, characterized in that, The target model is trained through the following steps: A first initial model is trained based on the first training data to obtain the first model, wherein the first training data includes the topological structure information of the sample job chain, the node table data of each sample job node in the sample job chain, and the runtime of each sample job node. The second initial model is trained based on the second training data to obtain the second model. The second training data includes the topological structure information of the sample job chain, the differentially processed table data of each sample job node in the sample job chain, and the runtime of each sample job node. The third initial model is trained based on the third training data to obtain the third model. The third training data includes the topology information of sample job chains of multiple different business types, the node table data of each sample job node in the sample job chains of multiple different business types, and the runtime of each sample job node in the sample job chains of multiple different business types. The target model is obtained based on the first model, the second model, and the third model.

4. The method according to claim 1, characterized in that, Constructing a directed acyclic graph based on the target predicted runtime of each target job node includes: Each target job node is treated as a graph node in the directed acyclic graph, and the target predicted runtime of each target job node is used as the weight of each graph node. The directed acyclic graph is constructed based on each graph node and its weight.

5. The method according to claim 1, characterized in that, The abnormal job nodes identified in the target job chain based on the directed acyclic graph include: Calculate the total time consumption of each path in the directed acyclic graph and identify the target path in the directed acyclic graph, wherein the target path is the path with the longest total time consumption among the first k sorted paths; Calculate the time difference between the predicted runtime and the actual runtime of each target job node in the target path; Calculate the difference between the resource utilization rate and the historical resource utilization rate for each target job node in the target path; The target job nodes whose duration difference is greater than the first preset difference threshold and / or whose resource utilization difference is greater than the second preset difference threshold are identified as the abnormal job nodes.

6. The method according to claim 5, characterized in that, The analysis report generated based on the directed acyclic graph and the abnormal job nodes for the target job chain includes: Based on the target path and the abnormal operation node, optimization suggestions are generated through the rule engine; Based on the target predicted runtime of each target job node, the target path, the abnormal job nodes, and the optimization suggestions, an analysis report for the target job chain is generated using a predefined analysis report template.

7. The method according to claim 1, characterized in that, After generating an analysis report for the target job chain based on the abnormal job nodes, the method further includes: Based on the analysis report, optimize the target workflow and monitor the optimization results of the optimized target workflow; The target model is adjusted based on the optimization results.

8. An automated analysis device for a work chain, characterized in that, include: The first determining unit is used to determine the target job chain to be analyzed, and to obtain the topology information of the target job chain and the node table data of each target job node in the target job chain. The input unit is used to input the topology information and the node table data of each target job node into the target model, and output the target predicted runtime of each target job node through the target model. The target model includes a first model, a second model and a third model, and the third model includes multiple sub-models for job nodes that process different business types. Construction unit, used to build a directed acyclic graph based on the target predicted runtime of each target job node; The identification unit is used to identify abnormal job nodes in the target job chain based on the directed acyclic graph. The generation unit is used to generate an analysis report of the target job chain based on the directed acyclic graph and the abnormal job nodes.

9. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, executes the automated analysis method for the job chain according to any one of claims 1 to 7.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the automated analysis method for the job chain according to any one of claims 1 to 7.

Images