Distributed timed task processing state detection method and device and storage medium

Abstract

The invention discloses a distributed timed task processing state detection method and device and a storage medium, and relates to the field of distribution.The method comprises the steps that resource occupation data of a target node in a parameter scheduling period is collected; obtaining the scheduling frequency of the N distributed timed tasks; converting the collected resource occupation data into frequency domain information; matching the frequency information of each frequency component in the frequency domain information with the scheduling frequency of the N distributed timed tasks to obtain a matching result; according to a matching result and phase information of the N distributed timed tasks, determining a processing state of the N distributed timed tasks by the target node, the phase information of the N distributed timed tasks including a task scheduling moment of each distributed timed task, the technical problem that in the prior art, the timeliness of state analysis is poor due to the fact that the processing state of the distributed timed task is detected through log analysis is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of distribution, in particular to a processing state detection method and device for distributed timing tasks and a storage medium. BACKGROUND

[0002] In the field of finance, e-commerce, Internet of Things and other distributed systems, the scheduling and execution state monitoring of distributed timing tasks is a key link to ensure stable operation of the system.

[0003] In the prior art, a distributed holographic monitoring method is usually used to locate the consumption delay phase difference of the consumer group, the execution cycle consistency of the distributed timing task and other problems. However, the existing monitoring method needs to extract various logs from the distributed timing task corresponding nodes with upstream and downstream calling relationship for analysis to see whether there are a series of problems such as cascading delay caused by distributed transaction coordination exception or message queue backlog. However, log analysis often takes a lot of time due to the analysis of various massive data, and if only resource occupation information is used, there is too little information available for further analysis, and manual log checking is required. The existing monitoring and alarm method does not have good predictability and cannot quickly and accurately warn of the occurrence of problems. SUMMARY

[0004] The embodiments of the present application provide a processing state detection method and device for distributed timing tasks and a storage medium to at least solve the technical problem that the state analysis is not timely due to the detection of the processing state of the distributed timing task by log analysis in the prior art.

[0005] According to an aspect of an embodiment of the present application, a processing state detection method for distributed timing tasks is provided, comprising: collecting resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1; obtaining scheduling frequencies of the N distributed timing tasks; converting the collected resource occupation data into frequency domain information; matching frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result; and determining a processing state of the target node for the N distributed timing tasks according to the matching result and phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks includes a task scheduling time of each distributed timing task.

[0006] Optionally, before collecting the resource occupation data of the target node in the parameter scheduling period, the method further comprises: detecting the task scheduling period of each distributed timing task; taking the distributed timing task with the longest task scheduling period as a target task according to the task scheduling period of each distributed timing task; and determining the parameter scheduling period according to the task scheduling period of the target task, wherein the parameter scheduling period is greater than or equal to the task scheduling period of the target task.

[0007] Optionally, converting the collected resource occupation data into frequency domain information comprises: performing Fourier transform on the collected resource occupation data according to the time sequence information of the data collection process to obtain the frequency domain information, wherein each frequency component in the frequency domain information comprises amplitude information and phase information corresponding to the frequency component.

[0008] Optionally, matching the frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result comprises: detecting whether the frequency information of each frequency component in the frequency domain information is the same as the frequency information of each scheduling frequency of the N distributed timing tasks; when it is detected that the frequency information of an i th frequency component is the same as the frequency information of a j th scheduling frequency, it is determined that the i th frequency component and the j th scheduling frequency match successfully, and task information corresponding to the i th frequency component and the j th scheduling frequency is grouped into a reference pair, wherein i and j are both integers greater than or equal to 1; and when it is detected that the frequency information of the i th frequency component is not the same as the frequency information of the j th scheduling frequency, it is determined that the i th frequency component and the j th scheduling frequency match unsuccessfully.

[0009] Optionally, determining the processing state of the N distributed timing tasks of the target node according to the matching result and the phase information of the N distributed timing tasks comprises: after obtaining at least one reference pair according to the matching result, detecting the real-time difference of the reference pair, wherein the real-time difference of the reference pair is used to represent the phase difference value between the frequency component and the task information included in the reference pair; and determining the processing state of the N distributed timing tasks of the target node according to the real-time difference of the reference pair.

[0010] Optionally, determining the processing state of the target node on the N distributed timing tasks according to the real-time difference of the reference pair comprises: obtaining a standard deviation set for the reference pair; when detecting that a difference value between the real-time difference of the reference pair and the standard deviation is greater than a preset threshold, determining that the processing process of the target node on the N distributed timing tasks is in an abnormal state, and generating a warning information; when detecting that the difference value between the real-time difference of the reference pair and the standard deviation is less than or equal to the preset threshold, determining that the processing process of the target node on the N distributed timing tasks is in a normal state.

[0011] Optionally, detecting the real-time difference of the reference pair comprises: detecting a task scheduling time point corresponding to task information included in the reference pair, wherein the task scheduling time point comprises a scheduling start time point and a scheduling end time point; detecting a frequency information collection time point corresponding to a frequency component included in the reference pair, wherein the frequency information collection time point comprises a collection start time point and a collection end time point; and determining the real-time difference of the reference pair according to the task scheduling time point and the frequency information collection time point.

[0012] According to another aspect of the embodiments of the present application, a distributed timing task processing state detection device is further provided, comprising: an acquisition unit configured to acquire resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1; an obtaining unit configured to obtain scheduling frequencies of the N distributed timing tasks; a first processing unit configured to convert the acquired resource occupation data into frequency domain information; a matching unit configured to match frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result; and a second processing unit configured to determine a processing state of the target node on the N distributed timing tasks according to the matching result and phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks comprises a task scheduling time point of each distributed timing task.

[0013] According to another aspect of the embodiments of the present application, a computer readable storage medium is further provided, and the computer readable storage medium stores a computer program, wherein when the computer program runs, the computer readable storage medium causes a device where the computer readable storage medium is located to perform the distributed timing task processing state detection method.

[0014] According to another aspect of the embodiments of the present application, an electronic device is also provided, including one or more processors and a memory, the memory being configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the processing state detection method of the distributed timing task described above.

[0015] According to another aspect of the embodiments of the present application, a computer program product is also provided, including a computer program or instructions, which, when executed by a processor, implement the processing state detection method of the distributed timing task described above.

[0016] In the present application, first, resource occupation data of a target node in a parameter scheduling period is collected, wherein the target node is a computing node participating in processing N distributed timing tasks, N is an integer greater than 1, and the scheduling frequency of the N distributed timing tasks is obtained, then the collected resource occupation data is converted into frequency domain information, and the frequency information of each frequency component in the frequency domain information is matched with the scheduling frequency of the N distributed timing tasks to obtain a matching result, according to the matching result and the phase information of the N distributed timing tasks, the processing state of the N distributed timing tasks by the target node is determined, wherein the phase information of the N distributed timing tasks includes the task scheduling time of each distributed timing task.

[0017] From the above, it can be seen that by collecting resource occupation data and converting it into frequency domain information, the present application can monitor the execution state of the task in real time, discover abnormalities in time, reduce the delay of log analysis, and by frequency matching and phase information analysis, the execution state of the task can be accurately identified, the problem can be quickly located, and the accuracy of problem diagnosis is improved. By setting a reasonable threshold and standard deviation, false positives and false negatives can be effectively reduced, and the reliability of the system is improved. The present application can flexibly adapt to distributed systems of different scales through distributed holographic monitoring and frequency domain analysis.

[0018] In summary, the present application collects resource occupation data of a target node in a parameter scheduling period, obtains the scheduling frequency of a distributed timing task, converts the resource occupation data into frequency domain information, performs frequency matching, and finally determines the task processing state, and by frequency domain analysis and phase information matching, real-time monitoring and accurate analysis of the processing state of the distributed timing task are achieved, effectively solving the technical problem that the processing state of the distributed timing task is detected through log analysis in the prior art, resulting in poor timeliness of state analysis. BRIEF DESCRIPTION OF DRAWINGS

[0019] The accompanying drawings, which are included to provide a further understanding of the present application, form a part of the present application and illustrate the illustrative embodiments of the present application and the explanation of the present application, and do not constitute improper limitations on the present application. In the drawings:

[0020] Figure 1 is a flow diagram of a distributed timing task processing state detection method according to the prior art;

[0021] Figure 2 is a principle diagram of an optional distributed timing task processing state detection method according to an embodiment of the present application;

[0022] Figure 3 is a diagram of an optional distributed timing task processing state detection device according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the personnel in the technical field better understand the scheme of the present application, the technical scheme in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by the person skilled in the art without creative labor should belong to the scope of protection of the present application.

[0024] It should be noted that the terms "first", "second" and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product or device.

[0025] It should also be noted that the information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) collected by the present application are information and data authorized by the user or authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of related data comply with relevant laws, regulations and standards in relevant regions, necessary security measures are taken, do not violate public order and good customs, and provide corresponding operation portal for user to choose authorization or refusal. For example, the system and related users or institutions are provided with an interface, and before obtaining the relevant information, the interface needs to send a request to the aforementioned user or institution, and after receiving the consent information feedback from the aforementioned user or institution, the relevant information is obtained.

[0026] According to the embodiment of the present application, an embodiment of a distributed timing task processing state detection method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] According to the embodiment of the present application, a data processing system can be used as the execution subject of the distributed timing task processing state detection method of the embodiment of the present application, wherein the system can be a software system or a combination of software and hardware embedded system. Of course, the method execution subject in the embodiment of the present application can also be other forms of execution subject, such as device, equipment, etc. Those skilled in the art should know that the specific form of the method execution subject is not particularly limited in the present application.

[0028] According to the embodiment of the present application, an embodiment of a distributed timing task processing state detection method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] Figure 1 The distributed timing task processing state detection method according to the embodiment of the present application, as shown in Figure 1 includes the following steps:

[0030] Step S101, collecting resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1;

[0031] In step S101, the target node can refer to a computing node participating in processing multiple distributed timing tasks. In the parameter scheduling period, the resource occupation data of the target node is collected, for example, the resource occupation data of the target node can include the use of CPU, memory, disk, network and other resources and their changes over time. In most cases, the resource occupation data changes positively, that is, there is no case where the CPU occupation is extremely high and the memory occupation is very low. According to the requirements of different scenes for troubleshooting of the target node, one or two resource occupation conditions can be selected for collection, for example, if the target node has more network congestion, the corresponding network resource occupation condition can be collected.

[0032] The parameter scheduling period can be determined based on the scheduling periods of all upstream tasks. For example, the parameter scheduling period can refer to the maximum value of the scheduling periods of all upstream tasks. It can be set as a single parameter or flexibly varied according to the periods of all upstream tasks. For instance, if the period for task 1 is hourly, the scheduling period for task 2 is daily, and the scheduling period for task 3 is every half month, then the parameter scheduling period is half a month, retaining the continuous change data collected within these half months. The continuous collection period will be much shorter than the parameter scheduling period, such as once every 5 minutes. The period parameter depends on production experience.

[0033] This application embodiment continuously collects resource usage data, enabling real-time acquisition of node operating status and reducing the analysis lag caused by log generation and collection delays in traditional log analysis. The collected resource data can be directly related to distributed scheduled tasks, directly reflecting the impact of distributed scheduled task execution on node resources, rather than requiring the sifting of useful information from massive amounts of irrelevant logs as in traditional log analysis.

[0034] Step S102: Obtain the scheduling frequency of N distributed timed tasks;

[0035] In step S102, all upstream tasks related to the target node are reviewed, and the scheduling frequency of each distributed scheduled task is recorded, i.e., the time interval between the execution of the distributed scheduled task. For example, the scheduling frequency of distributed scheduled task 1 is once per hour, and the scheduling frequency of distributed scheduled task 2 is once per day, etc. Scheduling frequency is one of the fundamental characteristics of distributed scheduled tasks. Obtaining the scheduling frequency helps in subsequent analysis of the relationship between distributed scheduled tasks and resource consumption, providing a foundation for establishing a coupled model of the scheduling frequency of distributed scheduled tasks and the frequency of resource fluctuations. By clearly defining the scheduling frequency of each distributed scheduled task, the complex distributed task scheduling problem can be decomposed into sub-problems with clear periodic characteristics, reducing the blind analysis of large amounts of unordered logs in traditional log analysis.

[0036] Step S103: Convert the collected resource occupancy data into frequency domain information;

[0037] In step S103, the acquired resource occupancy time-series data is transformed from the time domain to the frequency domain using Fourier transform. For example, frequency domain information may include the frequency, amplitude, and phase information of each frequency component. This transformation can clearly reveal the periodic variation characteristics of resource occupancy.

[0038] Optionally, the frequency domain analysis can extract the periodic characteristics in the resource occupation data, which can be related to the scheduling frequency of the distributed timing task, thereby providing a clearer basis for subsequent matching and analysis. The frequency domain analysis can filter out random noise in the time domain data and highlight the main periodic signal, making the analysis result more accurate. Traditional log analysis is mainly based on time domain data, which is difficult to directly capture the periodic characteristics of task scheduling. Frequency domain analysis can analyze data from another dimension, reducing the limitations of traditional analysis methods.

[0039] In step S104, the frequency information of each frequency component in the frequency domain information is matched with the scheduling frequency of the N distributed timing tasks to obtain a matching result.

[0040] In step S104, the frequency components in the frequency domain information can be matched one by one with the scheduling frequency of the N distributed timing tasks. Optionally, if the frequency of a certain frequency component is consistent with the scheduling frequency of a certain distributed timing task, it is considered that the matching result is successful, and the frequency component and the corresponding distributed timing task information form a reference pair. Through frequency matching, the frequency component of resource occupation can be associated with a specific distributed timing task, thereby establishing a strong coupling relationship between the task scheduling frequency and the resource fluctuation frequency. This direct association makes problem positioning more explicit and reduces the analysis difficulty caused by unclear task and resource occupation relationship in traditional log analysis.

[0041] In step S105, the processing state of the target node to the N distributed timing tasks is determined according to the matching result and the phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks includes the task scheduling time of each distributed timing task.

[0042] In step S105, based on the matching result and the phase information of the distributed timing task, the real-time difference of the reference pair is calculated. Optionally, the phase information of the distributed timing task includes the start time and end time of the task scheduling time. For example, according to the task call list, the scheduling frequencies of all upstream tasks related to the target node are F1 to Fn, and the time points of the calls are recorded as phase information W1 to Wn. All upstream tasks refer to the most source initiator, and the task call list is a call chain. Except for the most source and the task of the target node itself, all other tasks on the chain are not included in the scope of the analysis, and the phase information includes the start time point and the end time point of the call. The real-time difference of the reference pair can refer to the difference between the phase information of the frequency component and the phase information of the task. According to the difference value between the real-time difference and the standard difference, it is judged whether the processing state of the target node to the distributed timing task is normal. If the difference value exceeds the preset threshold, it is considered that the processing state is abnormal, and a warning information is generated.

[0043] The embodiment of the application can discover the abnormality in the early stage of the problem, and reduce the delay problem of the early warning caused by the need for post-analysis of a large number of logs in the traditional log analysis, by monitoring the change of the phase difference in real time. The change of the phase difference can reflect the problems such as the clock of the task scheduling being out of synchronization, the abnormality of the distributed transaction coordination, and the like, and can accurately locate the root cause of the problem, while the traditional log analysis can only discover the surface phenomenon and is difficult to quickly locate the root cause of the problem. The embodiment of the application can realize the dynamic monitoring of the distributed timing task processing state, and can reflect the running state of the system in real time, while the traditional log analysis is usually static and is difficult to meet the real-time monitoring demand of the distributed system.

[0044] In summary, the embodiment of the application can realize the fast and accurate monitoring and early warning of the distributed timing task processing state, reduce the limitations of the traditional log analysis, improve the efficiency and accuracy of the distributed system monitoring, and effectively solve the problem of poor analysis timeliness caused by the log analysis for detecting the distributed timing task processing state in the prior art, by the cooperation operation of collecting the resource occupation data, obtaining the scheduling frequency of the distributed timing task, frequency domain conversion, frequency matching and phase difference analysis.

[0045] Among them, the distributed holographic monitoring is an intelligent monitoring technology combining distributed computing and holographic data acquisition and analysis, aiming to realize all-around, high-precision and real-time dynamic environmental perception and abnormality detection.

[0046] In an optional embodiment, before collecting the resource occupation data of the target node in the parameter scheduling period, the method further includes: first detecting the task scheduling period of each distributed timing task, then taking the distributed timing task with the longest task scheduling period as a target task according to the task scheduling period of each distributed timing task, and determining the parameter scheduling period according to the task scheduling period of the target task, wherein the parameter scheduling period is greater than or equal to the task scheduling period of the target task.

[0047] Optionally, detecting the scheduling period of each distributed timing task can refer to the execution frequency of the distributed timing task, and the scheduling period reflects the execution interval time of the distributed timing task, for example, task A is executed once per hour, task B is executed once per day, and task C is executed once per half month. By detecting the scheduling period, the execution characteristics of each task are determined, and a time reference is provided for subsequent data collection and analysis.

[0048] Optionally, among all distributed timing tasks, the task with the longest scheduling period is selected as the target task. For example, if the period of task A is 1 hour, the period of task B is 1 day, and the period of task C is 15 days, the scheduling period of task C is the longest, and task C is selected as the target task. Selecting the task with the longest period as the target task can enable the execution of all tasks to be covered within the parameter scheduling period, and can avoid missing the execution data of any task.

[0049] Optionally, the parameter scheduling period is set to the scheduling period of the target task, and the parameter scheduling period is greater than or equal to the task scheduling period of the target task. For example, if the scheduling period of the target task is 15 days, the parameter scheduling period can be set to 15 days or more. By setting the parameter scheduling period to the scheduling period of the longest task, the collected data can completely cover the execution of all tasks, while avoiding collecting too much irrelevant data. Resource index collection can follow the minimization principle, and differential privacy technology can be used to inject noise into time series data to ensure that user behavior trajectories cannot be inferred.

[0050] By selecting the task with the longest period as the target task, the parameter scheduling period can completely cover the execution of all tasks, avoid missing the execution data of some tasks due to a too short collection period, avoid collecting too much irrelevant data, reduce the burden of data storage and processing, and improve the efficiency of data collection. By setting the parameter scheduling period to the scheduling period of the longest task, the setting of the data collection period is simplified, complex period calculation and adjustment are avoided, and the operation of the monitoring system is more efficient. The parameter scheduling period can be flexibly adjusted according to the scheduling period of the target task, so that the monitoring system can adapt to distributed systems with different task scheduling frequencies, improving the versatility and scalability. By reasonably setting the parameter scheduling period, the data collected and analyzed during non-task execution is avoided, the noise data interference on the analysis result is reduced, and the analysis accuracy is improved. By defining the relationship between the task scheduling period and the parameter scheduling period, the association between the tasks and the resource occupation can be more clearly established, providing a more accurate basis for subsequent frequency domain analysis and problem positioning. The traditional log analysis method needs to process a large amount of log data, and the analysis process takes a long time, making it difficult to find problems in time. By optimizing the data collection period, the embodiments of the present application can collect and analyze data more quickly, improving the timeliness of monitoring. The traditional log analysis method lacks clear task scheduling period information, making it difficult to accurately associate task and resource occupation data, and easily leading to misjudgment. By defining the relationship between the task scheduling period and the parameter scheduling period, the present application can more accurately associate task and resource occupation data, improving the accuracy of monitoring.

[0051] In an optional embodiment, converting the collected resource occupation data into frequency domain information comprises: performing Fourier transform on the collected resource occupation data according to time sequence information of the data collection process to obtain the frequency domain information, wherein each frequency component in the frequency domain information comprises amplitude information and phase information corresponding to the frequency component.

[0052] Optionally, the collected resource occupation data is recorded in chronological order, that is, each data point has a corresponding time stamp, and the data points reflect the resource occupation of the target node at different time points. Fourier transform is a mathematical tool that can convert time domain signals into frequency domain signals. Through Fourier transform, the resource occupation data can be converted from the time domain to the frequency domain. In the frequency domain, the signal is decomposed into a combination of sine waves and / or cosine waves of different frequencies, and each frequency component has corresponding amplitude and phase information. The converted frequency domain information includes the frequency, amplitude and phase of each frequency component. Frequency represents the periodicity of the signal, amplitude represents the strength of the signal, and phase represents the relative time offset of the signal.

[0053] Optionally, the frequency component can refer to in the frequency domain, the signal is decomposed into multiple frequency components. Each frequency component corresponds to a specific frequency value. Amplitude information can refer to amplitude representing the strength of the frequency component, reflecting the importance of the frequency component in the signal. Phase information can refer to phase representing the phase offset of the frequency component relative to a reference point. Phase information is very important for analyzing the periodicity and synchronicity of the signal. The phase difference between frequency components can refer to the phase relationship between frequency components when the signal contains multiple frequency components. The phase relationship between these frequency components is very critical to the waveform characteristics of the signal. For example, two sine waves of the same frequency will add up to a larger amplitude if the phase is the same; if the phase difference is 180 degrees, they will cancel each other out.

[0054] The embodiments of the present application convert complex time domain signals into frequency domain signals through Fourier transform, which can clearly extract the periodicity characteristics of the signal. The periodicity characteristics are directly related to the scheduling frequency of the distributed timing task, thereby providing a clear basis for subsequent analysis. Frequency domain analysis decomposes the signal into multiple frequency components, each of which has clear frequency, amplitude and phase information, making the analysis more intuitive and simple, avoiding the direct processing of complex signals in traditional time domain analysis. Frequency domain analysis can filter out random noise in the time domain signal, highlighting the main periodic signal, making the analysis result more accurate and clearly reflecting the impact of task scheduling on resource occupation. Phase information is an important part of frequency domain analysis, which can accurately capture the synchronicity problem of task scheduling. For example, by analyzing the phase difference, problems such as task scheduling clock desynchronization and distributed transaction coordination exception can be quickly found.

[0055] The phase information in the frequency domain refers to, in frequency domain analysis, the phase difference of each frequency component of a signal relative to a reference point, which can be the time zero point. In signal processing, any time-domain signal can be converted to the frequency domain by Fourier transform and decomposed into a combination of sine and / or cosine waves of different frequencies. The phase represents the angular deviation of the frequency component relative to the reference phase.

[0056] Traditional log analysis is mainly based on time domain data, and it is difficult to directly capture the periodic characteristics of task scheduling. Frequency domain analysis analyzes data from another dimension and can more comprehensively reflect the running state of the system. Frequency domain analysis can monitor the frequency and phase changes of the signal in real time, thereby achieving dynamic monitoring and rapid warning of the distributed timing task processing state, reducing the limitations of traditional log analysis, and enabling problems to be discovered and processed in a timely manner. Frequency domain analysis converts time domain signals into frequency components, reducing the amount of data that needs to be analyzed. By focusing on the main frequency components, problems can be quickly located, avoiding the processing of a large amount of irrelevant data in traditional log analysis. Through the matching of frequency and phase information, the relationship between task scheduling and resource occupation can be quickly determined, thereby quickly locating the root cause of the problem. This greatly improves the efficiency of the monitoring system and reduces the problem troubleshooting time.

[0057] Optionally, referring to Figure 2 In this embodiment, a frequency domain analysis machine is added, which converts the resource occupation curve into frequency domain information through Fourier transform. The frequency information of each frequency component in the frequency domain information is extracted and matched with the scheduling frequency of all tasks. The frequency components and upstream task records that match are recorded as a reference pair.

[0058] The amplitude information in the frequency component cannot have an exact corresponding relationship with the amplitude of the so-called resource occupation curve due to various influencing factors between the upstream task and the called task of the target node, but the frequency and phase information have a clear corresponding relationship, so only the frequency and phase information in the frequency domain are extracted. The difference between the one-to-one corresponding frequency component and the phase information of the upstream task in the reference pair is taken as a standard deviation by the frequency domain analysis machine. After setting the standard deviation of each reference pair, the phase information of the extracted frequency component is continuously collected, Fourier transformed, and the phase information of the upstream task is calculated. The difference between the extracted phase information and the real-time difference is calculated.

[0059] The real-time difference includes two differences of the start time point and the end time point. For example, the time points of the calls of all tasks 1 to N upstream are recorded as phase information W1 to Wn, and W1 of task 1 includes two time point information of W1begin and W1end. In reference to the above, the phase information of the frequency component also includes two time point information, so that the standard difference and the real-time difference both contain the difference of the start time point and the difference of the end time point in the reference pair.

[0060] In an optional embodiment, the frequency information of each frequency component in the frequency domain information is matched with the scheduling frequencies of the N distributed timing tasks to obtain a matching result, including: detecting whether the frequency information of each frequency component in the frequency domain information is same as the frequency information of each scheduling frequency of the N distributed timing tasks; when it is detected that the frequency information of the ith frequency component is same as the frequency information of the jth scheduling frequency, it is determined that the ith frequency component and the jth scheduling frequency match successfully, and the task information corresponding to the ith frequency component and the jth scheduling frequency is combined into a reference pair, where i and j are both integers greater than or equal to 1; when it is detected that the frequency information of the ith frequency component is not same as the frequency information of the jth scheduling frequency, it is determined that the ith frequency component and the jth scheduling frequency match unsuccessfully.

[0061] Optionally, detecting whether the frequency component and the scheduling frequency are same can mean comparing the frequency value of each frequency component in the frequency domain information with the scheduling frequency of each distributed timing task one by one to determine whether they are same. Through the frequency matching, it is determined which frequency components are caused by the distributed timing tasks, so as to establish a direct association between the resource occupation and the tasks.

[0062] Optionally, if the frequency value of the ith frequency component is same as the scheduling frequency of the jth task, it is considered that the matching is successful. The frequency component matched successfully and the corresponding task information are combined into a reference pair for subsequent phase difference analysis. For example, the frequency component fi is same as the scheduling frequency Fj of the task j, and a reference pair (fi, task j, Fj) is formed. If the frequency value of the ith frequency component is not same as the scheduling frequency of any task, it is considered that the matching is unsuccessful. For example, the frequency component matched unsuccessfully is excluded from the subsequent analysis, and it is considered that these frequency components are caused by other factors, such as noise or non-task-related resource fluctuation.

[0063] The embodiments of the present application can directly associate the frequency component occupied by the resource with the specific distributed timing task through frequency matching, establish a strong coupling relationship between the task scheduling frequency and the resource fluctuation frequency, make the problem positioning more direct and clear, and effectively avoid the analysis difficulty caused by the unclear relationship between the task and the resource occupation in the traditional log analysis. For example, if the frequency component f1 matches the scheduling frequency F1 of task 1 successfully, it can be explicitly known that the execution of task 1 has a periodic impact on the resource occupation. The matching successful reference pair provides an explicit benchmark for subsequent phase difference analysis. By analyzing the phase information of the frequency component and the phase information of the task in the reference pair, it can be further judged whether the execution state of the task is normal. For example, the difference value, i.e. real-time difference, between the phase information of the frequency component and the phase information of the task in the reference pair can be calculated to determine whether the task has problems such as scheduling clock out of synchronization and distributed transaction coordination exception. The matching successful reference pair can accurately reflect the impact of task scheduling on resource occupation, so that subsequent analysis is more focused, only the matching successful frequency component and task need to be concerned, and the analysis of a large amount of irrelevant data is avoided, thereby improving the efficiency of problem positioning. For example, if only a small number of frequency components match the task successfully, subsequent analysis only needs to focus on these reference pairs, and there is no need to process all frequency components, thereby greatly reducing the workload of analysis. By strictly screening the frequency components related to the task scheduling frequency, false judgments caused by noise or interference signals can be avoided. For example, the frequency components that fail to match are considered as noise signals irrelevant to task scheduling, and can be excluded from subsequent analysis, thereby reducing the possibility of false positives and false negatives.

[0064] In an optional embodiment, according to the matching result and the phase information of the N distributed timing tasks, the processing state of the target node to the N distributed timing tasks is determined, including: after obtaining at least one reference pair according to the matching result, detecting the real-time difference of the reference pair, wherein the real-time difference of the reference pair is used to represent the phase difference value between the frequency component and the task information included in the reference pair; and determining the processing state of the target node to the N distributed timing tasks according to the real-time difference of the reference pair.

[0065] Optionally, for each matching successful reference pair, the difference value, i.e. real-time difference, between the phase information of the frequency component and the phase information of the task is detected. The real-time difference includes the difference between the task scheduling time and the frequency component collection time. For example, the scheduling start time of the task is Tbegin, and the collection start time of the frequency component is Fbegin, then the real-time difference is ΔTbegin=Tbegin-Fbegin. The real-time difference of the end time ΔTend can also be calculated. The real-time difference reflects the deviation between the actual execution of the current task and the expected execution, and is used to determine whether the task is executed as planned.

[0066] Optionally, according to the size and variation trend of the real-time difference, it is determined whether the processing state of the target node to the distributed timing task is normal. If the absolute value or the change rate of the real-time difference exceeds a preset threshold, it indicates that the execution state of the task is abnormal, for example, the scheduling clock is out of synchronization or the distributed transaction coordination is abnormal. If the real-time difference is within the preset threshold range, it indicates that the execution state of the task is normal. Through the monitoring of the real-time difference, the execution state of the task can be quickly and accurately determined, and abnormalities can be found and handled in a timely manner.

[0067] The embodiments of the present application can monitor the execution state of the task in real time by detecting the real-time difference, and find the deviation of the task execution in a timely manner. Compared with the traditional log analysis, problems can be found faster, and the analysis lag problem caused by log generation and collection delay in log analysis can be avoided. If the real-time difference of the task suddenly increases, an alarm can be triggered immediately to prompt that the task may have scheduling problems. The real-time difference can accurately reflect the phase deviation of the task execution, for example, whether the task is started on time, whether it is completed on time, etc. This accuracy makes problem positioning more accurate and can quickly identify problems such as the scheduling clock being out of synchronization and the distributed transaction coordination being abnormal. If the start time real-time difference of the task continues to increase, it may be that the scheduling clock is out of synchronization; if the end time real-time difference continues to increase, it may be a distributed lock preemption problem. The detection of the real-time difference is dynamic and can adapt to changes in the execution state of the task. Even if the execution time of the task drifts slightly, such changes can be captured in a timely manner, thereby achieving dynamic monitoring. If the scheduling frequency of the task changes, the monitoring of the real-time difference can be adjusted in a timely manner, which can improve the accuracy of monitoring. By setting a reasonable threshold, false positives can be effectively reduced. Only when the real-time difference is indeed outside the normal range, an alarm will be triggered, thereby improving the reliability of the system.

[0068] In an optional embodiment, the processing state of the target node to the N distributed timing tasks is determined according to the real-time difference of the reference pair, comprising: obtaining a standard deviation set for the reference pair; when it is detected that the difference value between the real-time difference of the reference pair and the standard deviation is greater than a preset threshold, determining that the processing process of the target node to the N distributed timing tasks is in an abnormal state, and generating a warning information; when it is detected that the difference value between the real-time difference of the reference pair and the standard deviation is less than or equal to the preset threshold, determining that the processing process of the target node to the N distributed timing tasks is in a normal state.

[0069] Optionally, the standard deviation can refer to the expected difference value between the phase information of the frequency component in the reference pair and the phase information of the task under normal circumstances. The standard deviation can be obtained through historical data analysis or manually set according to system design experience. The standard deviation provides a clear benchmark for the judgment of the real-time difference, making the abnormality detection more scientific and objective.

[0070] Optionally, a difference value between the real-time difference and the standard difference of the reference pair is calculated. The difference value is usually represented by an absolute value, for example, Δ = |real-time difference - standard difference|. By quantifying the difference value, it can be more intuitive to determine whether the real-time difference is outside the normal range.

[0071] Optionally, according to the comparison result of the difference value and the preset threshold value, the processing state is determined: if the difference value is greater than the preset threshold value, it means that the deviation between the real-time difference and the standard difference is too large, and the processing process of the task is in an abnormal state, and the system generates a warning information. If the difference value is less than or equal to the preset threshold value, it means that the real-time difference is within the normal range, and the processing process of the task is in a normal state. By setting a reasonable threshold value, false positives can be effectively reduced, and abnormal states can be discovered and processed in time.

[0072] The embodiments of the present application can discover abnormalities and generate warning information in the early stage of problems by monitoring the difference value between the real-time difference and the standard difference, which is faster than the traditional log analysis method, and can reduce the impact of problems on system performance. If the difference value between the real-time difference and the standard difference suddenly increases, the system can immediately trigger an alarm to prompt that there may be a scheduling problem with the task. By setting a reasonable threshold value, false positives can be effectively reduced. Only when the difference value is indeed outside the normal range, an alarm will be triggered, thereby improving the reliability of the system. By analyzing historical data to set a reasonable threshold value, false positives caused by normal fluctuations can be avoided. By clear judgment criteria, the monitoring system can automatically determine the state, reducing manual intervention and improving monitoring efficiency. The system can automatically determine the task state and generate warning information without manual checking of logs.

[0073] Optionally, referring to Figure 2 , the frequency domain analysis machine monitors the change of the difference between the start time point and the end time point in the real-time difference and the standard difference of each reference pair, and triggers an alarm when the change of the real-time difference relative to the standard difference exceeds a threshold value. Different situations are triggered for changes, triggering a system adaptive calibration process or reminding manual calibration. The number of system adaptive calibration steps can be the same as the number of target nodes. The situation where the change exceeds the threshold value is designed differently for different scenarios, for example, the change of the real-time difference relative to the standard difference, whether it is the start time point or the end time point, is constantly widening, which means that there is a problem of unsynchronized scheduling clock or cascading delay, such as network time protocol clock deviation, heartbeat error of distributed scheduling framework. Only the difference value related to the end time point is constantly widening, and the difference value related to the start time point is unchanged, which means that distributed lock preemption occurs, and the design and coding of the corresponding distributed transaction coordination mechanism need to be checked. The size of the threshold value depends on production experience.

[0074] For example, one of the typical scenarios in distributed systems: when multiple services rely on periodic tasks, such as timed synchronization, batch jobs, if there is a slight difference in the scheduling frequency, such as scheduling based on different server clocks, the phase difference of delay fluctuations will gradually accumulate, eventually causing task execution misalignment, such as the response trough period of the upstream service coinciding with the request peak period of the downstream service, causing performance bottlenecks. What the adaptive calibration process is depends on what the problem is, such as different scheduling clocks, the system automatically recalibrates the clocks of each node, and such as the heartbeat error of the distributed scheduling framework, the system updates the scheduling algorithm, such as using a more strict synchronization mechanism or dynamically adjusting the task execution window, which has mature practices and will not be repeated here.

[0075] Optionally, the frequency domain analysis machine can accurately locate the corresponding task call link as the focus of problem investigation according to the corresponding reference pair of the corresponding upstream task. For example, in some scenarios, the change of the real-time difference of only a few reference pairs exceeds the threshold, that is, only the call link of the corresponding upstream task to the target node of these reference pairs has a related mechanism problem, and checking the parameter settings, related logs, distributed design codes, etc. on the link can be checked. Through frequency domain analysis and frequency domain phase difference dynamic modeling, phase drift at a specific frequency is captured, and problems such as network time protocol clock deviation, distributed scheduling framework heartbeat error, and distributed lock preemption, distributed transaction coordination mechanism are more quickly and accurately located. Unlike the time domain backtracking mechanism of traditional log analysis, the present application is based on frequency domain phase difference dynamic modeling to realize explicit feature extraction of implicit task dependency, breaking through the limitations of artificial experience.

[0076] In an optional embodiment, detecting the real-time difference of the reference pair comprises: detecting a task scheduling time point corresponding to the task information included in the reference pair, wherein the task scheduling time point comprises a scheduling start time point and a scheduling end time point; detecting a frequency information collection time point corresponding to the frequency component included in the reference pair, wherein the frequency information collection time point comprises a collection start time point and a collection end time point; and determining the real-time difference of the reference pair according to the task scheduling time point and the frequency information collection time point.

[0077] Optionally, detecting the task scheduling time point can refer to obtaining the scheduling time point of the task information in the reference pair, including the scheduling start time point and the scheduling end time point of the task, which can be obtained through the task scheduling system or log record, and can clearly define the actual execution time range of the task, providing an accurate time reference for subsequent phase difference analysis.

[0078] Optionally, the frequency information collection time point can refer to a collection time point of acquiring a reference frequency component, including a collection start time point and a collection end time point, which can be acquired through the record of the frequency domain analysis system, and can clearly define the time range in which the frequency component is recorded, thereby providing an accurate time reference for subsequent phase difference analysis.

[0079] Optionally, determining the real-time difference can refer to calculating the difference between the task scheduling time point and the frequency information collection time point, i.e., the real-time difference. Specifically, the real-time difference can include the difference between the task scheduling start time point and the frequency component collection start time point, and the difference between the task scheduling end time point and the frequency component collection end time point. Through the real-time difference, the deviation between the actual execution of the task and the expected execution can be quantified, so as to determine whether the execution state of the task is normal.

[0080] By calculating the real-time difference, the embodiments of the present application can monitor the execution state of the task in real time and discover the deviation of the task execution in a timely manner. Compared with traditional log analysis, this can discover problems faster and avoid the analysis lag problem caused by log generation and collection delay in log analysis. If the real-time difference of the task suddenly increases, an alarm can be triggered immediately to prompt that there may be a scheduling problem with the task. The real-time difference can accurately reflect the phase deviation of the task execution, such as whether the task is started on time or completed on time. This accuracy makes problem positioning more accurate and can quickly identify problems such as task scheduling clock desynchronization and distributed transaction coordination exception. If the start time real-time difference of the task continuously increases, it may be that the scheduling clock is desynchronized; if the end time real-time difference continuously increases, it may be a distributed lock preemption problem. The calculation of the real-time difference is dynamic and can adapt to changes in the execution state of the task. Even if the execution time of the task drifts slightly, this change can be captured in a timely manner, thereby achieving dynamic monitoring. If the scheduling frequency of the task changes, the monitoring of the real-time difference can be adjusted in a timely manner, thereby improving the accuracy of monitoring. By setting a reasonable threshold, false positives can be effectively reduced. Only when the real-time difference is indeed outside the normal range will an alarm be triggered, thereby improving the reliability of the system. By analyzing historical data to set a reasonable threshold, false positives caused by normal fluctuations can be avoided.

[0081] According to another aspect of the embodiments of the present application, a processing state detection device for distributed timing tasks is also provided, as shown in Figure 3 The device includes a collection unit 31, an acquisition unit 32, a first processing unit 33, a matching unit 34, and a second processing unit 35.

[0082] The collecting unit 31 is configured to collect resource occupation data of a target node in a parameter scheduling period, the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1. The obtaining unit 32 is configured to obtain scheduling frequencies of the N distributed timing tasks. The first processing unit 33 is configured to convert the collected resource occupation data into frequency domain information. The matching unit 34 is configured to match frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result. The second processing unit 35 is configured to determine a processing state of the target node for the N distributed timing tasks according to the matching result and phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks includes a task scheduling time of each distributed timing task.

[0083] Optionally, the processing state detection apparatus of the distributed timing task further includes a detection unit configured to detect a task scheduling period of each distributed timing task, a third processing unit configured to determine a distributed timing task with a longest task scheduling period as a target task according to the task scheduling period of each distributed timing task, and a fourth processing unit configured to determine a parameter scheduling period according to the task scheduling period of the target task, wherein the parameter scheduling period is greater than or equal to the task scheduling period of the target task.

[0084] Optionally, the first processing unit 33 includes a processing subunit configured to perform Fourier transform on the collected resource occupation data to obtain the frequency domain information according to time sequence information of a data collection process, wherein each frequency component in the frequency domain information includes amplitude information and phase information corresponding to the frequency component.

[0085] Optionally, the second processing unit 35 includes a first detection subunit configured to detect whether frequency information of each frequency component in the frequency domain information is same as frequency information of each scheduling frequency of the N distributed timing tasks, a matching subunit configured to determine that an ith frequency component and a jth scheduling frequency match successfully when it is detected that the frequency information of the ith frequency component is same as the frequency information of the jth scheduling frequency, and configured to group task information corresponding to the ith frequency component and the jth scheduling frequency to form a reference pair, wherein i and j are both integers greater than or equal to 1, and a first determination subunit configured to determine that the ith frequency component and the jth scheduling frequency match unsuccessfully when it is detected that the frequency information of the ith frequency component is not same as the frequency information of the jth scheduling frequency.

[0086] Optionally, the second processing unit 35 further comprises: a second detection sub-unit, configured to detect a real-time difference of the reference pair after obtaining the at least one reference pair according to the matching result, wherein the real-time difference of the reference pair is used to represent a phase difference value between the frequency component included in the reference pair and the task information; and a second determination sub-unit, configured to determine the processing state of the N distributed timing tasks of the target node according to the real-time difference of the reference pair.

[0087] Optionally, the second determination sub-unit comprises: an acquisition module, configured to acquire a standard deviation set for the reference pair; a first processing module, configured to determine that the processing process of the N distributed timing tasks of the target node is in an abnormal state and generate an early warning information when detecting that a difference value between the real-time difference of the reference pair and the standard deviation is greater than a preset threshold; and a second processing module, configured to determine that the processing process of the N distributed timing tasks of the target node is in a normal state when detecting that the difference value between the real-time difference of the reference pair and the standard deviation is less than or equal to the preset threshold.

[0088] Optionally, the second detection sub-unit comprises: a first detection module, configured to detect a task scheduling time corresponding to the task information included in the reference pair, wherein the task scheduling time comprises a scheduling start time and a scheduling end time; a second detection module, configured to detect a frequency information acquisition time corresponding to the frequency component included in the reference pair, wherein the frequency information acquisition time comprises an acquisition start time and an acquisition end time; and a determination module, configured to determine the real-time difference of the reference pair according to the task scheduling time and the frequency information acquisition time.

[0089] According to another aspect of the embodiments of the present application, a computer readable storage medium is further provided, wherein the computer readable storage medium stores a computer program, and when the computer program runs, the computer readable storage medium causes the device where the computer readable storage medium is located to perform the processing state detection method of the distributed timing task.

[0090] According to another aspect of the embodiments of the present application, an electronic device is further provided, wherein the electronic device comprises one or more processors and a memory, and the memory is configured to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors perform the processing state detection method of the distributed timing task.

[0091] According to another aspect of the embodiments of the present application, a computer program product is further provided, wherein the computer program product comprises a computer program or instructions, and when the computer program or instructions are executed by a processor, the processing state detection method of the distributed timing task is implemented.

[0092] The serial numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0093] In the above-described embodiments of the present application, the description of each embodiment focuses on different aspects, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0094] In several embodiments provided in the present application, it should be understood that the disclosed technical content can be implemented by other manners. Among them, the above-described device embodiments are only schematic, for example, the division of the units can be a logical function division, and in actual implementation, there can be another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units or modules shown or discussed can be indirect coupling or communication connection through some interfaces, units or modules, which can be electrical or other forms.

[0095] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed to a plurality of units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0096] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The above integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0097] The integrated unit, if realized in the form of software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part of the prior art that makes a contribution or the whole or part of the technical solutions can be embodied in the form of software product, which is stored in a storage medium and includes a plurality of instructions for making a computer device (which can be a personal computer, a server or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes various media that can store program codes, such as U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk, etc.

[0098] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principle of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be regarded as the protection scope of the present application.

Claims

1. A method for detecting the processing status of distributed timing tasks, characterized in that, The method comprises: collecting resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1; obtaining scheduling frequencies of the N distributed timing tasks; converting the collected resource occupation data into frequency domain information; matching frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result; determining a processing state of the target node for the N distributed timing tasks according to the matching result and phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks comprises a task scheduling time of each distributed timing task.

2. The method of claim 1, wherein, Before collecting the resource occupation data of the target node in the parameter scheduling period, the method further comprises: detecting a task scheduling period of each distributed timing task; taking a distributed timing task with the longest task scheduling period as a target task according to the task scheduling period of each distributed timing task; determining the parameter scheduling period according to the task scheduling period of the target task, wherein the parameter scheduling period is greater than or equal to the task scheduling period of the target task.

3. The method of claim 1, wherein, Converting the collected resource occupation data into frequency domain information comprises: performing Fourier transform on the collected resource occupation data according to time sequence information of a data collection process to obtain the frequency domain information, wherein each frequency component in the frequency domain information comprises amplitude information and phase information corresponding to the frequency component.

4. The method of claim 1, wherein, Matching the frequency information of each frequency component in the frequency domain information with the scheduling frequencies of the N distributed timing tasks to obtain a matching result comprises: detecting whether the frequency information of each frequency component in the frequency domain information is the same as the frequency information of each scheduling frequency of the N distributed timing tasks; when it is detected that the frequency information of an i th frequency component is the same as the frequency information of a j th scheduling frequency, determining that the i th frequency component and the j th scheduling frequency match successfully, and grouping task information corresponding to the i th frequency component and the j th scheduling frequency to form a reference pair, wherein i and j are both integers greater than or equal to 1; when it is detected that the frequency information of the i th frequency component is different from the frequency information of the j th scheduling frequency, determining that the i th frequency component and the j th scheduling frequency match unsuccessfully.

5. The method of claim 4, wherein, Determining the processing state of the target node for the N distributed timing tasks according to the matching result and the phase information of the N distributed timing tasks comprises: after obtaining at least one reference pair according to the matching result, detecting a real-time difference of the reference pair, wherein the real-time difference of the reference pair is used to represent a phase difference value between the frequency component and the task information included in the reference pair; determining the processing state of the target node for the N distributed timing tasks according to the real-time difference of the reference pair.

6. The method of claim 5, wherein, Determining the processing state of the target node for the N distributed timing tasks according to the real-time difference of the reference pair comprises: obtaining a standard deviation set for the reference pair; When the difference between the real-time difference of the reference pair and the standard difference is greater than a preset threshold, it is determined that the target node is in an abnormal state for processing the N distributed timing tasks, and a warning message is generated. When the difference between the real-time difference of the reference pair and the standard difference is less than or equal to the preset threshold, it is determined that the target node is in a normal state for processing the N distributed timing tasks.

7. The method of claim 5, wherein, Detecting the real-time difference of the reference pair includes: Detecting a task scheduling time corresponding to task information included in the reference pair, wherein the task scheduling time includes a scheduling start time and a scheduling end time; Detecting a frequency information collection time corresponding to a frequency component included in the reference pair, wherein the frequency information collection time includes a collection start time and a collection end time; Determining the real-time difference of the reference pair according to the task scheduling time and the frequency information collection time.

8. A processing status detection device for a distributed timed task, characterized in that, Comprise: The acquisition unit is used for acquiring resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1; The acquisition unit is used for acquiring resource occupation data of a target node in a parameter scheduling period, wherein the target node is a computing node participating in processing N distributed timing tasks, and N is an integer greater than 1; The first processing unit is used for converting the acquired resource occupation data into frequency domain information; The matching unit is used for matching frequency information of each frequency component in the frequency domain information with scheduling frequencies of the N distributed timing tasks to obtain a matching result; The second processing unit is used for determining a processing state of the target node for the N distributed timing tasks according to the matching result and phase information of the N distributed timing tasks, wherein the phase information of the N distributed timing tasks includes a task scheduling time of each distributed timing task.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, wherein when the computer program runs, the computer readable storage medium makes the device execute the distributed timing task processing state detection method in any one of claims 1 to 7.

10. An electronic device, comprising: Comprise one or more processors and memories, the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the distributed timing task processing state detection method in any one of claims 1 to 7.

11. A computer program product, characterised in that, Comprise computer programs or instructions, which realize the distributed timing task processing state detection method in any one of claims 1 to 7 when executed by a processor.

Images