Server fault diagnosis method and device, computer storage medium and electronic equipment

Abstract

The invention discloses a server fault diagnosis method and device, a computer storage medium and electronic equipment. The method comprises the steps of obtaining data information of a plurality of feature indexes of a server and a preset normal threshold range of each feature index; determining the feature deviation degree of the corresponding feature index according to the data information of the feature index and a corresponding preset normal threshold range; determining a feature comprehensive evaluation value according to the feature deviation degrees of the plurality of feature indexes and the corresponding weight coefficients; and performing fault diagnosis on the server according to the characteristic comprehensive evaluation value. Therefore, according to the method, feature diagnosis is carried out on the server through the multiple feature indexes, the fault diagnosis precision is improved, and the operation stability of the server is improved.

Description

Technical Field

[0001] The present application relates to the technical field of server fault diagnosis, and in particular to a server fault diagnosis method, a server fault diagnosis device, a computer-readable storage medium, and an electronic device. Background Art

[0002] In today's digital age, servers play a vital role in various information technology systems. The stable operation of servers is directly related to business continuity, data security, and user experience.

[0003] In the related art, the server is diagnosed for faults based on the CPU (Central Processing Unit) usage or memory usage of the server, but the fault diagnosis accuracy of this method is low, which reduces the operation stability of the server. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems in the related art to a certain extent. To this end, the first purpose of the present application is to propose a server fault diagnosis method, which performs feature diagnosis on the server through multiple feature indicators, improves the fault diagnosis accuracy, and improves the operation stability of the server.

[0005] The second objective of the present application is to provide a server fault diagnosis device.

[0006] The third object of the present application is to provide a computer-readable storage medium.

[0007] The fourth objective of the present application is to provide an electronic device.

[0008] To achieve the above-mentioned objectives, the first aspect of the present application proposes a server fault diagnosis method, which includes: obtaining data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator; determining the characteristic deviation of the corresponding characteristic indicator based on the data information of the characteristic indicator and the corresponding preset normal threshold range; determining a characteristic comprehensive evaluation value based on the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients; and performing server fault diagnosis based on the characteristic comprehensive evaluation value.

[0009] According to the server fault diagnosis method of the embodiment of the present application, firstly, data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator are obtained, characteristic deviation of the corresponding characteristic indicator is determined according to the data information of the characteristic indicator and the corresponding preset normal threshold range, and characteristic comprehensive evaluation value is determined according to the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients, and the server is fault diagnosed according to the characteristic comprehensive evaluation value. Thus, the method performs characteristic diagnosis on the server through multiple characteristic indicators, improves the accuracy of fault diagnosis, and provides guarantee for the stable operation of the server.

[0010] In addition, the server fault diagnosis method according to the above embodiment of the present application may also have the following additional technical features:

[0011] According to an embodiment of the present application, the characteristic deviation of the corresponding characteristic indicator is determined according to the data information of each characteristic indicator and the corresponding preset normal threshold range, including: when the data information of the characteristic indicator is within the preset normal threshold range, determining the characteristic deviation of the corresponding characteristic indicator to be zero; when the data information of the characteristic indicator exceeds the upper limit value of the corresponding preset normal threshold range, respectively obtaining a first difference between the data information and the upper limit value of the preset normal threshold range and a second difference between the upper limit value and the lower limit value of the preset normal threshold range, and determining the characteristic deviation of the corresponding characteristic indicator according to the ratio of the first difference to the second difference; when the data information of the characteristic indicator is lower than the lower limit value of the preset normal threshold range, respectively obtaining a third difference between the lower limit value of the preset normal threshold range and the data information and a second difference between the upper limit value and the lower limit value of the preset normal range, and determining the characteristic deviation of the corresponding characteristic indicator according to the ratio of the third difference to the second difference.

[0012] According to one embodiment of the present application, a feature comprehensive evaluation value is determined based on the feature deviations of multiple feature indicators and their corresponding weight coefficients, including: obtaining a fault evaluation value of the corresponding feature indicator based on the product of the feature deviations of the feature indicators and their corresponding weight coefficients; obtaining a feature comprehensive evaluation value based on the sum of the fault evaluation values ​​of multiple feature indicators.

[0013] According to one embodiment of the present application, fault diagnosis of the server is performed based on the comprehensive evaluation value of the features, including: when the comprehensive evaluation value of the features is greater than or equal to a preset fault threshold, determining that the server is in a high-risk state for fault; when the comprehensive evaluation value of the features is less than the preset fault threshold, determining that the server is in a normal or low-risk state for fault.

[0014] According to one embodiment of the present application, the server fault diagnosis method also includes: obtaining multiple historical feature comprehensive evaluation values ​​within a preset time period; determining a fault prediction coefficient sequence within the preset time period based on the multiple historical feature comprehensive evaluation values; and predicting server faults based on numerical changes in the fault prediction coefficient sequence.

[0015] According to an embodiment of the present application, the fault prediction coefficient sequence includes multiple fault prediction coefficients, and the fault prediction coefficients in the fault prediction coefficient sequence within a preset time period are calculated by the following formula:

[0016]

[0017] Among them, P k represents the kth fault prediction coefficient in the fault prediction coefficient sequence, E t represents the comprehensive evaluation value of the tth historical feature within the preset time period, The average value of the comprehensive evaluation values ​​of multiple historical features within the preset time period, E t-k It represents the comprehensive evaluation value of the tkth historical feature within the preset time period.

[0018] According to one embodiment of the present application, fault prediction of the server is performed based on the numerical change of the fault prediction coefficient sequence, including: when the numerical value of the fault prediction coefficient sequence shows an increasing trend, determining that the failure risk of the server is increased; when the numerical value of the fault prediction coefficient sequence shows a decreasing trend and stops changing when reaching a preset numerical value, determining that the failure risk of the server is reduced.

[0019] To achieve the above-mentioned objectives, the second aspect of the present application proposes a server fault diagnosis device, which includes: an acquisition module, used to obtain data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator; a first determination module, used to determine the characteristic deviation of the corresponding characteristic indicator based on the data information of the characteristic indicator and the corresponding preset normal threshold range; a second determination module, used to determine the characteristic comprehensive evaluation value based on the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients; a fault diagnosis module, used to diagnose server faults based on the characteristic comprehensive evaluation value.

[0020] According to the server fault diagnosis device of the embodiment of the present application, the acquisition module acquires data information of multiple characteristic indicators of the server and the preset normal threshold range of each characteristic indicator, the first determination module determines the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range, the second determination module determines the characteristic comprehensive evaluation value according to the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients, and the fault diagnosis module performs fault diagnosis on the server according to the characteristic comprehensive evaluation value. Thus, the device performs characteristic diagnosis on the server through multiple characteristic indicators, improves the fault diagnosis accuracy, and provides guarantee for the stable operation of the server.

[0021] To achieve the above-mentioned purpose, the third aspect of the present application proposes a computer-readable storage medium on which a server fault diagnosis program is stored. When the server fault diagnosis program is executed by a processor, the above-mentioned server fault diagnosis method is implemented.

[0022] According to the computer-readable storage medium of the embodiment of the present application, the above-mentioned server fault diagnosis method is implemented when the server fault diagnosis program stored thereon is executed by the processor. Based on the above-mentioned server fault diagnosis method, feature diagnosis of the server is performed through multiple feature indicators, thereby improving the accuracy of fault diagnosis and providing guarantee for the stable operation of the server.

[0023] To achieve the above-mentioned objectives, the fourth aspect embodiment of the present application proposes an electronic device, including a memory, a processor, and a server fault diagnosis program stored in the memory and executable on the processor. When the processor executes the server fault diagnosis program, the above-mentioned server fault diagnosis method is implemented.

[0024] According to the electronic device of the embodiment of the present application, when the server fault diagnosis program is executed by the processor, the above-mentioned server fault diagnosis method is implemented. Based on the above-mentioned server fault diagnosis method, feature diagnosis of the server is performed through multiple feature indicators, which improves the fault diagnosis accuracy and provides a guarantee for the stable operation of the server.

[0025] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of a server fault diagnosis method according to an embodiment of the present application;

[0027] Figure 2 is a flow chart of a server fault diagnosis method according to a specific embodiment of the present application;

[0028] Figure 3A flowchart of server failure prediction according to a specific embodiment of the present application;

[0029] Figure 4 A connection diagram of a server fault diagnosis device according to an embodiment of the present application;

[0030] Figure 5 Schematic diagram of a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0031] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0032] The following describes the server fault diagnosis method, server fault diagnosis device, computer-readable storage medium and electronic device proposed in the embodiments of the present application with reference to the accompanying drawings.

[0033] Figure 1 The present invention is a flowchart of a server fault diagnosis method according to an embodiment of the present application.

[0034] like Figure 1 As shown, the server fault diagnosis method of the embodiment of the present application may include:

[0035] S1, obtaining data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator;

[0036] S2, determining the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range;

[0037] S3, determining a comprehensive feature evaluation value according to the feature deviations of multiple feature indicators and their corresponding weight coefficients;

[0038] S4, diagnose server faults based on the comprehensive evaluation value of the features.

[0039] Specifically, multiple characteristic indicators can be set based on actual conditions, and multiple representative characteristic indicators can be determined by combing through various information sources of the server. For example, multiple characteristic indicators may include multiple dimensions such as hardware characteristic information, software characteristic information, sel (System Event Log) log information, sensor status information, performance indicator information, and system log information of the server. Through the selection of multi-dimensional characteristic indicators, the operating status of the server can be fully reflected, and the failure to be missed due to the limitations of a single feature can be avoided, which greatly improves the comprehensiveness and accuracy of fault diagnosis. For example, sel log information can capture abnormal events at the hardware level in a timely manner; sensor status information can monitor changes in the physical environment inside the server in real time; performance indicator information intuitively shows the operating efficiency of the server; system log information helps to discover hidden faults at the software level.

[0040] This embodiment has a high degree of flexibility and scalability in the selection of characteristic indicators. In addition to covering common sel log information, sensor status information, performance indicator information, and system log information, it also allows the inclusion of other information related to server fault diagnosis and prediction. This enables the method to adapt to the needs of continuous development and innovation of server technology, effectively respond to the challenges of different failure modes, and continue to maintain the efficiency and accuracy of fault diagnosis. For example, with the widespread application of server virtualization technology, newly added characteristic indicators related to virtual resource allocation and utilization can be included in the analysis framework of this application, providing strong support for server operation and maintenance in a virtualized environment.

[0041] In addition, each characteristic indicator sets a corresponding preset normal threshold range. Specifically, the preset normal threshold range can be set for each characteristic indicator by analyzing the historical data of the server under normal operation and industry standards and specifications. This provides a clear and scientific basis for subsequent fault diagnosis and prediction, and can quickly and accurately identify whether the characteristic indicator deviates from the normal range, effectively reducing the occurrence of misjudgment, thereby significantly improving the efficiency and reliability of fault diagnosis. The following is an example of setting the preset normal threshold range for the characteristic indicator:

[0042] 1) Hardware feature information

[0043] The preset normal threshold range of the CPU (Central Processing Unit) temperature indicator: According to the CPU model and the heat dissipation design of the server, set the normal range of the CPU temperature as the preset normal threshold range of the CPU temperature. For example, the temperature of a normal server CPU should be between 40℃-60℃ in idle state, and should not exceed 80℃ in full load state, so the preset normal threshold range of the CPU temperature indicator is set to [40℃,80℃].

[0044] Preset normal threshold range of memory frequency indicator: According to the memory specifications and CPU model support, determine the normal range of memory frequency as the preset normal threshold range of memory frequency. For example, if the normal frequency of a certain memory fluctuates within the range of 2400MHz±100MHz, then the preset normal threshold range of the memory frequency indicator is [2300MHz, 2500MHz].

[0045] The preset normal threshold range of the ECC (Error Checking and Correcting) times indicator, the ECC times include the memory ECC times and the PCIe (Peripheral Component Interconnect Express, a high-speed serial computer expansion bus standard) link ECC times. Then the preset normal threshold range of the ECC times indicator can be set with reference to the ECC alarm times of the memory and PCIe link. For example, when the ECC times reach 5000, a fault alarm will be issued, that is, when the ECC times reach 5000, it indicates that a fault has occurred. Then the method can choose to issue a pre-alarm at 80% of the time and make a prediction in advance, that is, the preset normal threshold range of the ECC times is [0,4000]. In addition, in addition to fault judgment based on the ECC times, parameters such as duration, intensity, and frequency can also be introduced for fault prediction. For example, in the ECC monitoring system of the server, if there is a frequent increase in the ECC times within a period of time, from several times a month to dozens of times a day, this may mean that some storage units in the memory module have begun to fail, and further inspection or replacement of the memory is required to avoid memory loss.

[0046] The preset normal threshold range of the hard disk read and write speed index: refer to the technical specifications and performance indicators of the hard disk to determine its normal read and write speed range, which is used as the preset normal threshold range of the hard disk read and write speed. For example, for a mechanical hard disk, the preset normal threshold range of the hard disk read speed is 80MB / s-120MB / s, and the preset normal threshold range of the hard disk write speed is 60MB / s-100MB / s; for a solid-state hard disk, the preset normal threshold range of the hard disk read speed is 300MB / s-500MB / s, and the preset normal threshold range of the hard disk write speed is 200MB / s-400MB / s.

[0047] The preset normal threshold range of the hard disk's smart (Self-Monitoring Analysis and Reporting Technology) parameter indicators can be set by referring to the preset thresholds provided by the hard disk manufacturer.

[0048] Preset normal threshold range of network card bandwidth indicator: set the normal utilization range of network card bandwidth according to the nominal bandwidth of the network card and the actual situation of the network environment as its preset normal threshold range. For example, for a 10Gbps network card, under normal business load, the bandwidth utilization should be between 20%-60%, that is, the sum of inbound and outbound traffic should not exceed 6Gbps. At this time, the preset normal threshold range can be set to 0-6Gbps.

[0049] The preset normal threshold range of the PCIe link speed indicator: Set the range of its normal link rate according to the theoretical value of the PCIe device channel as its preset normal threshold range. For example, for a PCIe3.0 device, the theoretical value of its single-channel (x1) link speed is 8GT / s, so the preset normal threshold range of the PCIe link speed is set to 6GT / s-8GT / s. The specific range value can also be adjusted based on historical data analysis and business scenarios.

[0050] 2) Software feature information

[0051] The preset normal threshold range of the operating system resource usage indicator: monitor the CPU and memory usage of the operating system during normal operation, and set its preset normal threshold range. In general, the CPU usage of the operating system itself should be less than 10% when idle and should not exceed 30% when processing regular tasks; the memory usage should be determined based on the total server memory and the number of installed services. For example, for a server with 32GB of memory, the memory occupied by the operating system and basic services should be between 4GB and 8GB, and the remaining memory can be used by other applications. Therefore, the preset normal threshold range of the operating system resource usage indicator is 4GB-8GB.

[0052] The preset normal threshold range of the application software response time indicator: For key application software running on the server, such as database management systems and Web (World Wide Web) server software, set the normal range of its response time. For example, the query response time of the database should be less than 10 milliseconds for simple queries and should not exceed 100 milliseconds for complex queries, so the preset normal threshold range can be set to [0, 100ms]; the page request response time of the server should be between 50 milliseconds and 200 milliseconds, so the preset normal threshold range can be set to [0, 200ms]. The specific value can be adjusted according to the server performance and the load of the application.

[0053] 3) Log feature range

[0054] Preset normal threshold range of SEL event frequency indicator: According to the historical operation data and hardware reliability indicators of the server, determine the normal occurrence frequency of different types of SEL events to determine the preset normal threshold range of SEL event frequency indicators. For example, the number of hardware error events per month should not exceed 5 times; the number of power failure events per quarter should not exceed 1 time; the number of overtemperature events per week should not exceed 1 time. If these frequency ranges are exceeded, it may indicate that the server has a potential failure risk.

[0055] The preset normal threshold range of the system log error and warning frequency indicators: The normal frequency of occurrence of different types of error and warning information in the statistical system log is used as the preset normal threshold range. For example, the "insufficient disk space" warning should not appear more than 3 times per month; the "software crash" error should not appear more than 1 time per quarter. Exceeding these frequencies may indicate that there is a problem with the server's software or hardware configuration, which requires further investigation and optimization.

[0056] 3) Sensor characteristic information

[0057] Preset normal threshold range of fan speed indicator: According to the design specifications and heat dissipation requirements of the server fan, determine the normal range of the fan speed to determine its preset normal threshold range. For example, the normal speed range of the server fan is between 2000RPM-4000RPM. When the speed is lower than 1500RPM, it may not provide enough cooling air volume, resulting in an increase in hardware temperature; when the speed is higher than 4500RPM, it may indicate that the fan is faulty or the internal temperature of the server is too high, and the fan is in an overloaded state. Therefore, 2000RPM-4000RPM can be used as the preset normal threshold range of the fan speed indicator.

[0058] The preset normal threshold range of voltage fluctuation index: According to the technical parameters and stability requirements of the server power supply, the normal fluctuation range of voltage is set to determine its preset normal threshold range. For example, the fluctuation range of +12V voltage should be within ±5%, and the preset normal threshold range of voltage fluctuation index is determined to be 11.4V-12.6V; the fluctuation range of +5V voltage should be within ±3%, and the preset normal threshold range of voltage fluctuation is determined to be 4.85V-5.15V; the fluctuation range of +3.3V voltage should be within ±3%, and the preset normal threshold range of voltage fluctuation is determined to be 3.20V-3.40V. If the voltage fluctuation exceeds these ranges, it may affect the normal operation of the server hardware and even cause hardware damage.

[0059] After determining the data information of each characteristic indicator and the corresponding preset normal threshold range, the characteristic deviation of each characteristic indicator is calculated according to the data information of each characteristic indicator and the corresponding preset normal threshold range. For example, for the CPU temperature indicator, its preset normal threshold range is [40°C, 80°C]. Assuming that the current actual temperature obtained is 85°C, the current actual temperature is within the normal threshold range, no characteristic deviation occurs, and the characteristic deviation is 0; if the current actual temperature obtained is 95°C, the current actual temperature is outside the normal threshold range, and a deviation of 5°C occurs, the characteristic deviation can be calculated by calculating the ratio of 5°C to the preset normal threshold range, or the characteristic deviation can be calculated based on the preset deviation parameter, without specific limitation.

[0060] After calculating the characteristic deviation of each characteristic indicator, the characteristic comprehensive evaluation value is calculated according to the weight coefficient corresponding to each characteristic indicator. It can be understood that the sum of the weight coefficients of multiple characteristic indicators is equal to 1. The weight coefficient can be determined by combining expert experience and data-driven, where expert experience can fully consider the special needs and key factors of the server in different application scenarios, industry standards and previous operation and maintenance experience; data-driven uses big data analysis technology to mine the inherent correlation between characteristic indicators and faults from massive server operation data, so as to determine a more scientific and reasonable weight distribution. This comprehensive weight determination method enables the fault diagnosis and prediction method to better adapt to the diverse needs of different types of servers, further improving the versatility and effectiveness of the method. For example, in the server operation and maintenance of the financial industry, experts may appropriately increase the weights of characteristic indicators related to data encryption and storage based on the extremely high requirements of financial transactions for data security and stability; and data-driven analysis can further optimize these weights based on the fault cases in the actual operation data, making the fault diagnosis and prediction results more accurate.

[0061] Therefore, this embodiment performs fault diagnosis on the server according to multiple characteristic indicators, improves the comprehensiveness and accuracy of fault diagnosis, and effectively ensures the stable operation of the server.

[0062] In one embodiment of the present application, the characteristic deviation of the corresponding characteristic indicator is determined based on the data information of each characteristic indicator and the corresponding preset normal threshold range, including: when the data information of the characteristic indicator is within the preset normal threshold range, the characteristic deviation of the corresponding characteristic indicator is determined to be zero; when the data information of the characteristic indicator exceeds the upper limit value of the corresponding preset normal threshold range, the first difference between the data information and the upper limit value of the preset normal threshold range and the second difference between the upper limit value and the lower limit value of the preset normal threshold range are respectively obtained, and the characteristic deviation of the corresponding characteristic indicator is determined according to the ratio of the first difference to the second difference; when the data information of the characteristic indicator is lower than the lower limit value of the preset normal threshold range, the third difference between the lower limit value of the preset normal threshold range and the data information and the second difference between the upper limit value and the lower limit value of the preset normal range are respectively obtained, and the characteristic deviation of the corresponding characteristic indicator is determined according to the ratio of the third difference to the second difference.

[0063] Specifically, after obtaining the real-time data of multiple characteristic indicators of the server, i.e., data information, the characteristic deviation is introduced to quantify the degree to which each characteristic indicator deviates from the preset normal threshold range. The characteristic deviation of each characteristic indicator can be calculated by the following formula:

[0064]

[0065] Among them, D i Indicates the characteristic deviation of the i-th characteristic index; F i Represents the data information of the i-th characteristic index; [L i , U i ] represents the preset normal threshold range of the i-th characteristic index, where L i represents the lower limit of the preset normal threshold range of the i-th characteristic indicator; U i Indicates the upper limit value of the preset normal threshold range of the i-th characteristic indicator.

[0066] For example, for the CPU temperature indicator, assuming that the current actual temperature is 85°C and its preset normal threshold range is [40°C, 80°C], the characteristic deviation of the CPU temperature indicator is:

[0067]

[0068] This indicates that the CPU temperature is outside the normal range, with a characteristic deviation of 0.125, indicating that the server may have a heat dissipation problem and requires further attention.

[0069] For the memory usage indicator, if the current memory usage is 40% and the preset normal threshold range is [20%, 60%], the characteristic deviation of the memory usage indicator is:

[0070] D i =0

[0071] This means that the memory usage indicator is within the normal range and there is no abnormality in the server's memory usage.

[0072] Therefore, in this embodiment, when the data information of the characteristic indicator is within the preset normal threshold range, the characteristic deviation D i A value of 0 indicates that the server is in a normal state on this characteristic indicator; when the data information of this characteristic indicator exceeds the preset normal threshold range, the characteristic deviation D i Calculation is performed based on the excess ratio, with positive values ​​indicating an excess of the upper limit and negative values ​​indicating a loss below the lower limit. In this way, the contribution of each characteristic indicator to fault diagnosis can be intuitively reflected, further improving the accuracy of fault diagnosis and providing strong data support for subsequent comprehensive evaluation.

[0073] In one embodiment of the present application, a feature comprehensive evaluation value is determined based on the feature deviations of multiple feature indicators and their corresponding weight coefficients, including: obtaining a fault evaluation value of the corresponding feature indicator based on the product of the feature deviations of the feature indicators and their corresponding weight coefficients; obtaining a feature comprehensive evaluation value based on the sum of the fault evaluation values ​​of multiple feature indicators.

[0074] That is to say, after obtaining the characteristic deviations of multiple characteristic indicators, the characteristic comprehensive evaluation value is calculated by the following formula:

[0075]

[0076] Among them, E represents the comprehensive evaluation value of the feature, n represents the total number of multiple feature indicators, and w i represents the weight coefficient of the i-th characteristic index, D i represents the characteristic deviation of the i-th characteristic index, (w i ×D i ) represents the fault assessment value of the i-th characteristic index.

[0077] For example, assuming that 10 characteristic indicators are determined, their weights are as follows: the weight coefficient w1 of the CPU temperature indicator is 0.15, the weight coefficient w2 of the memory usage indicator is 0.1, the weight coefficient w3 of the hard disk read and write speed indicator is 0.1, the weight coefficient w4 of the network card bandwidth utilization indicator is 0.05, the weight coefficient w5 of the operating system CPU usage indicator is 0.1, the weight coefficient w6 of the application software response time indicator is 0.1, the weight coefficient w7 of the SEL event frequency indicator is 0.1, the weight coefficient w8 of the system log error frequency indicator is 0.1, the weight coefficient w9 of the fan speed indicator is 0.05, and the weight coefficient w10 of the voltage fluctuation indicator is 0. 10 =0.05.

[0078] Assume that at a certain moment, the index deviations of the characteristic indicators are as follows: the characteristic deviation D1 of the CPU temperature indicator is 0.125, the characteristic deviation D2 of the memory usage indicator is 0, the characteristic deviation D3 of the hard disk read and write speed indicator is 0, the characteristic deviation D4 of the network card bandwidth utilization indicator is 0, the characteristic deviation D5 of the operating system CPU usage indicator is 0.05, the characteristic deviation D6 of the application software response time indicator is 0.2, the characteristic deviation D7 of the SEL event frequency indicator is 0, the characteristic deviation D8 of the system log error frequency indicator is 0, the characteristic deviation D9 of the fan speed indicator is 0 (fan speed), and the characteristic deviation D1 of the voltage fluctuation indicator is 0. 10 =0, then the comprehensive fault assessment value is:

[0079] E=0.15×0.125+0.1×0+0.1×0+0.05×0+0.1×0.05+0.1×0.2+0.1×0+0.05×0+0.05×0

[0080] =0.18775+0+0+0+0.005+0.02+0+0+0

[0081] =0.043775

[0082] This embodiment can determine the weight coefficient of each fault indicator based on the intrinsic correlation between the characteristic indicators and the faults, so as to adapt it to the diverse needs of different types of servers. At the same time, it ensures the data validity of the comprehensive fault evaluation value, provides strong data support for subsequent comprehensive evaluation, and makes fault diagnosis more accurate.

[0083] In one embodiment of the present application, a server fault diagnosis is performed based on a feature comprehensive evaluation value, including: when the feature comprehensive evaluation value is greater than or equal to a preset fault threshold, determining that the server is in a high-risk fault state; when the feature comprehensive evaluation value is less than the preset fault threshold, determining that the server is in a normal or low-risk fault state.

[0084] That is to say, by setting a preset fault threshold, the calculated comprehensive fault assessment value is compared with the preset fault threshold. When the comprehensive fault assessment value is greater than the preset fault threshold, the server is judged to have a high fault risk; when the comprehensive fault assessment value is less than or equal to the preset fault threshold, the server is judged to be in a normal or low-risk state. When it is determined that the server has a high fault risk, the operation and maintenance personnel can immediately start the troubleshooting process, quickly locate the fault source based on the deviation information of each characteristic indicator, and repair it in time to avoid business interruption.

[0085] The fault judgment mechanism adopted in this embodiment makes the fault diagnosis result clear at a glance, which is convenient for operation and maintenance personnel to take corresponding measures in time, effectively reduces the losses caused by untimely fault discovery, and significantly improves the efficiency and reliability of server operation and maintenance.

[0086] Further, the preset fault threshold in this embodiment can be dynamically and flexibly adjusted according to the actual operating environment of the server, historical fault data and actual needs. For example, during the peak period of server business load, the fault threshold is appropriately increased to avoid false alarms caused by normal business fluctuations. Specifically, one or more characteristic parameters such as CPU utilization, memory utilization, network bandwidth utilization, disk I / O, etc. can be used to determine whether the server has entered the peak period of business load. For example, if the CPU utilization rate has been maintained at more than 95% in a certain period of time, it is considered that the server is in the peak period of business load; if the memory utilization rate has been maintained at more than 80% in a certain period of time, it is considered that the server is in the peak period of business load. In addition, during critical business periods (such as major meetings, securities trading hours, etc.) with extremely high requirements for server stability, the fault threshold is lowered to increase the sensitivity of fault detection and improve the operational reliability of the server. This dynamic adjustment mechanism enables the fault diagnosis and prediction method to better adapt to the complex and changeable operating conditions of the server, further improves the adaptability and reliability of the method, and provides a more accurate guarantee for the stable operation of the server.

[0087] In one embodiment of the present application, the server fault diagnosis method also includes: obtaining multiple comprehensive evaluation values ​​of historical features within a preset time period; determining a fault prediction coefficient sequence within the preset time period based on the multiple comprehensive evaluation values ​​of historical features; and predicting server faults based on numerical changes in the fault prediction coefficient sequence.

[0088] Specifically, the preset time period can be set according to actual needs, such as one day, one week, etc. The fault prediction coefficient sequence of the preset time period is obtained by performing data analysis on the comprehensive evaluation values ​​of historical features within the preset time period. The fault prediction coefficient sequence includes multiple fault prediction coefficients. For example, the fault prediction coefficient sequence is established by obtaining the correlation coefficients of multiple comprehensive evaluation values ​​of historical features. Then, by observing the numerical changes of multiple fault prediction coefficients, the server is predicted to have a fault. For example, when the fault prediction coefficient shows an increasing trend, it is considered that the failure risk of the server is gradually increasing; when the fault prediction coefficient shows a numerically stable trend, it is considered that the server is tending to be stable.

[0089] This embodiment effectively captures the potential rules and trends in the historical feature comprehensive evaluation values ​​based on data analysis, improves the reliability of the prediction results, and further combines the application of multiple feature indicators in the historical feature comprehensive evaluation values, so that it can comprehensively and accurately realize the server fault prediction and effectively ensure the stable operation of the server. In addition, the server fault prediction results can also be further combined with the current feature comprehensive evaluation value for the final fault diagnosis output. For example, in the case of a predicted server failure risk gradually increasing, if the current feature comprehensive evaluation value is less than the preset fault threshold, the server failure risk can be further judged in combination with the degree of proximity between the current feature comprehensive evaluation value and the preset fault threshold. For example, when the difference between the current feature comprehensive evaluation value and the preset fault threshold is less than the preset difference, the server is considered to be at a high failure risk; otherwise, the server is considered to be at a low failure risk, so that the operation and maintenance personnel can perform server maintenance in advance.

[0090] Therefore, in order to achieve accurate fault prediction, multiple historical data of the server are collected and sorted. In addition to the comprehensive evaluation value of historical features, historical data can also include historical data information of feature indicators, without specific restrictions. Historical data provides sufficient data resources for subsequent in-depth data analysis, which can dig out the changing rules of the server's operating status and the potential trend of fault occurrence, and provide a solid data foundation for fault prediction, thereby greatly improving the reliability and foresight of fault prediction. For example, by analyzing historical data from the past few months or even years, it can be found whether the server is more likely to fail in a specific season or during business peak periods, providing a strong basis for taking preventive measures in advance.

[0091] In addition, the collection of historical data includes data storage and management and data collection frequency and time span, as follows:

[0092] For the data storage and management part: first, establish a special database or data storage system to store the historical data of the server. You can use the preset data storage structure and indexing technology to ensure the fast writing, query and retrieval of data. For example, use a relational database management system to store data and use timestamps as indexes to facilitate chronological query and analysis of data. At the same time, regularly back up and archive historical data to prevent data loss. Specifically, you can use a combination of local backup and off-site backup to ensure data security and reliability. For example, back up local data once a day and back up data to an off-site storage server every week.

[0093] Regarding the frequency and time span of data collection: First, determine the appropriate frequency of data collection, which must ensure that changes in the server's operating status can be captured, but not too frequent, resulting in excessive data volume that affects storage and analysis efficiency. For example, for comprehensive fault assessment values ​​and key performance indicators (such as CPU usage, memory usage, etc.), data can be collected every 1 minute; for SEL events and system log information, they can be recorded in real time when the event occurs, and summarized and organized regularly (such as every hour). In addition, it is necessary to accumulate a sufficiently long time span of historical data so that the changing patterns of the server in different time periods, different business loads, and operating environments can be analyzed. For example, collect historical data from the past 3 months, or collect historical data from the past 6 months. For some server systems with extremely high stability requirements, historical data for a longer period of time (such as more than 1 year) can be collected.

[0094] In one embodiment of the present application, the fault prediction coefficient sequence includes multiple fault prediction coefficients, and the fault prediction coefficients in the fault prediction coefficient sequence within a preset time period are calculated by the following formula:

[0095]

[0096] Among them, P k represents the kth fault prediction coefficient in the fault prediction coefficient sequence, E t It represents the comprehensive evaluation value of the t-th historical feature in the preset time period. The average value of the comprehensive evaluation values ​​of multiple historical features within a preset time period, E t-k It represents the comprehensive evaluation value of the tkth historical feature within the preset time period.

[0097] Specifically, in formula (3), k represents the lag order, and the maximum value of k can be set according to the actual situation. Assuming that the total number of fault prediction coefficients in the fault prediction coefficient sequence is set to 6, that is, the maximum lag order is 6, then based on formula (3), the fault prediction coefficients P1, P2, P3, P4, P5 and P6 are calculated respectively to obtain the fault prediction coefficient sequence, and the server fault is predicted according to the numerical changes of the fault prediction coefficients P1, P2, P3, P4, P5 and P6.

[0098] This embodiment calculates the fault prediction coefficient sequence within the preset time period through formula (3). Through this sequence, the correlation of the comprehensive evaluation value of historical characteristics under different time lags can be deeply revealed, and the potential periodicity, trend and other complex laws in the data can be accurately captured, which provides a key technical means for fault prediction and effectively improves the accuracy and scientificity of fault prediction. For example, if the fault prediction coefficient shows obvious periodic changes, it indicates that the server may have a periodic fault risk. The operation and maintenance personnel can prepare in advance and perform system maintenance and optimization before the high-incidence period of faults arrives to reduce the probability of faults.

[0099] In one embodiment of the present application, fault prediction of the server is performed based on the numerical change of the fault prediction coefficient sequence, including: when the numerical value of the fault prediction coefficient sequence shows an increasing trend, determining that the failure risk of the server is increased; when the numerical value of the fault prediction coefficient sequence shows a decreasing trend and stops changing when reaching a preset numerical value, determining that the failure risk of the server is reduced.

[0100] Specifically, through in-depth analysis of the fault prediction coefficient sequence, observing its numerical size, numerical changes, periodicity, trend and other characteristics, the changing trend of the future comprehensive fault evaluation value is predicted.

[0101] If the value of the failure prediction coefficient sequence shows a gradually rising trend, it may indicate that the risk of server failure will gradually increase in the future. Operation and maintenance personnel should strengthen the monitoring frequency of the server, prepare in advance the necessary hardware spare parts and software repair tools, such as spare hard disks, memory modules, and related software patches, and formulate emergency plans so that they can respond quickly when a failure occurs and reduce business interruption time.

[0102] If the value of the fault prediction coefficient sequence shows a trend of gradually decreasing and approaching the preset value (for example, zero), it indicates that the server operation status tends to be stable and the risk of failure is reduced. The operation and maintenance personnel can appropriately adjust the monitoring strategy and devote more energy to the maintenance of other servers or systems, but still need to maintain a certain period of regular inspections to ensure the continuous and stable operation of the server and prevent new potential faults. Among them, the preset value is a reference stability value set according to the actual situation.

[0103] As a specific embodiment of the present application, Figure 2 As shown, the server fault diagnosis method of this embodiment includes:

[0104] S101, obtaining data information of multiple characteristic indicators of a server and a preset normal threshold range of each characteristic indicator.

[0105] S102, based on the data information of the characteristic indicator and the corresponding preset normal threshold range based on the formula

[0106] S103, calculating the comprehensive evaluation value of the feature

[0107] S104, determining whether the feature comprehensive evaluation value E is greater than or equal to a preset fault threshold T. If so, executing step S105; if not, executing step S106.

[0108] S105: Determine that the server is in a high-risk state for failure.

[0109] At this point, the operation and maintenance personnel can immediately start the troubleshooting process, quickly locate the source of the fault based on the deviation information of each characteristic indicator, and perform repairs in a timely manner to avoid business interruption.

[0110] S106, determining that the server is in a normal state or a low-risk state of failure.

[0111] In addition, the feature comprehensive evaluation value calculated based on the above method is stored for fault risk prediction, so that the server fault assessment can be performed in combination with the risk prediction result, such as Figure 3 As shown, the fault prediction method may include the following steps:

[0112] S201, obtaining comprehensive evaluation values ​​of multiple historical features within a preset time period.

[0113] S202, passed A plurality of fault prediction coefficients are calculated to construct a fault prediction coefficient sequence.

[0114] S203, judging whether the value of the fault prediction coefficient sequence presents an increasing trend. If yes, executing step S204; if no, executing step S205.

[0115] S204: Determine that the failure risk of the server increases.

[0116] At this time, operation and maintenance personnel should increase the frequency of server monitoring, prepare in advance the necessary hardware spare parts and software repair tools, such as spare hard disks, memory modules, and related software patches, and develop emergency plans so that they can respond quickly when a failure occurs and reduce business interruption time.

[0117] S205, judging whether the value of the fault prediction coefficient sequence shows a decreasing trend and stops changing when reaching a preset value. If so, executing step S206; if not, executing step S207.

[0118] S206: Determine that the failure risk of the server is reduced. At this point, the operation and maintenance personnel can appropriately adjust the monitoring strategy and devote more energy to the maintenance of other servers or systems.

[0119] S207, re-predicting server failure based on multiple historical feature comprehensive evaluation values ​​within another preset time period.

[0120] Therefore, this embodiment can bring the following beneficial effects:

[0121] 1. Improve fault prediction accuracy: Through autocorrelation analysis of the comprehensive evaluation values ​​of historical features, the potential regularities and periodic changes in server operation data can be accurately captured. Compared with the fault diagnosis and prediction methods based on real-time monitoring or simple threshold judgment in related technologies, it can more accurately predict the time and possibility of fault occurrence, reduce false alarms and missed alarms, and provide more reliable decision-making basis for operation and maintenance personnel.

[0122] 2. Realize early fault warning: Based on the trend analysis of the fault prediction coefficient, early signs of faults can be discovered in advance, and warnings can be issued before the faults have a serious impact, so that operation and maintenance personnel have enough time to take preventive measures, such as backing up data in advance, preparing spare parts, optimizing system configuration, etc., thereby effectively reducing the impact of faults on the business and improving the availability and stability of the server.

[0123] 3. Optimize the allocation of operation and maintenance resources: Based on the fault prediction results, operation and maintenance personnel can reasonably arrange working time and resources, strengthen monitoring and maintenance efforts during the high-incidence period of faults, and appropriately reduce unnecessary inspections and maintenance work during the low-incidence period of faults, so as to achieve efficient use of operation and maintenance resources and reduce operation and maintenance costs.

[0124] 4. Improve operation and maintenance efficiency: Based on the above method, operation and maintenance personnel can quickly locate components or links that may fail, reduce the time and workload of troubleshooting, improve the efficiency and quality of operation and maintenance work, shorten server downtime, and ensure business continuity.

[0125] 5. Enhance system scalability: This method has certain versatility and scalability. It is not only applicable to the current server system, but also can be applied to fault diagnosis and prediction of other similar complex information systems to a certain extent, providing strong support for the overall IT (Information Technology) infrastructure management of the enterprise, and has a broad application prospect and value.

[0126] In summary, according to the server fault diagnosis method of the embodiment of the present application, firstly, the data information of multiple characteristic indicators of the server and the preset normal threshold range of each characteristic indicator are obtained, the characteristic deviation of the corresponding characteristic indicator is determined according to the data information of the characteristic indicator and the corresponding preset normal threshold range, and the characteristic comprehensive evaluation value is determined according to the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients, and the server is fault diagnosed according to the characteristic comprehensive evaluation value. Therefore, the method performs characteristic diagnosis on the server through multiple characteristic indicators, improves the fault diagnosis accuracy, and ensures the operation stability of the server.

[0127] Corresponding to the above embodiment, the present application also proposes a server fault diagnosis device.

[0128] like Figure 4 As shown, the server fault diagnosis device of the embodiment of the present application may include: an acquisition module 10, a first determination module 20, a second determination module 30 and a fault diagnosis module 40.

[0129] Among them, the acquisition module 10 is used to obtain data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator. The first determination module 20 is used to determine the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range. The second determination module 30 is used to determine the characteristic comprehensive evaluation value according to the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients. The fault diagnosis module 40 is used to perform fault diagnosis on the server according to the characteristic comprehensive evaluation value.

[0130] According to one embodiment of the present application, the first determination module 20 determines the characteristic deviation of the corresponding characteristic indicator based on the data information of each characteristic indicator and the corresponding preset normal threshold range, and is specifically used for: when the data information of the characteristic indicator is within the preset normal threshold range, determining that the characteristic deviation of the corresponding characteristic indicator is zero; when the data information of the characteristic indicator exceeds the upper limit value of the corresponding preset normal threshold range, respectively obtaining the first difference between the data information and the upper limit value of the preset normal threshold range and the second difference between the upper limit value and the lower limit value of the preset normal threshold range, and determining the characteristic deviation of the corresponding characteristic indicator according to the ratio of the first difference to the second difference; when the data information of the characteristic indicator is lower than the lower limit value of the preset normal threshold range, respectively obtaining the third difference between the lower limit value of the preset normal threshold range and the data information and the second difference between the upper limit value and the lower limit value of the preset normal range, and determining the characteristic deviation of the corresponding characteristic indicator according to the ratio of the third difference to the second difference.

[0131] According to one embodiment of the present application, the second determination module 30 determines the feature comprehensive evaluation value based on the feature deviations of multiple feature indicators and their corresponding weight coefficients, and is specifically used to: obtain the fault evaluation value of the corresponding feature indicator based on the product of the feature deviations of the feature indicators and their corresponding weight coefficients; obtain the feature comprehensive evaluation value based on the sum of the fault evaluation values ​​of multiple feature indicators.

[0132] According to one embodiment of the present application, the fault diagnosis module 40 performs fault diagnosis on the server based on the feature comprehensive evaluation value, specifically for: when the feature comprehensive evaluation value is greater than or equal to a preset fault threshold, determining that the server is in a high-risk state of fault; when the feature comprehensive evaluation value is less than the preset fault threshold, determining that the server is in a normal or low-risk state of fault.

[0133] According to one embodiment of the present application, the fault diagnosis module 40 is also used to: obtain multiple comprehensive evaluation values ​​of historical features within a preset time period; determine a fault prediction coefficient sequence within a preset time period based on multiple comprehensive evaluation values ​​of historical features; and predict faults on the server based on numerical changes in the fault prediction coefficient sequence.

[0134] According to an embodiment of the present application, the fault prediction coefficient sequence includes multiple fault prediction coefficients, and the fault diagnosis module 40 calculates the fault prediction coefficients in the fault prediction coefficient sequence within a preset time period by the following formula:

[0135]

[0136] Among them, P k represents the kth fault prediction coefficient in the fault prediction coefficient sequence, E t It represents the comprehensive evaluation value of the t-th historical feature in the preset time period. The average value of the comprehensive evaluation values ​​of multiple historical features within a preset time period, E t-k It represents the comprehensive evaluation value of the tkth historical feature within the preset time period.

[0137] According to one embodiment of the present application, the fault diagnosis module 40 predicts faults of the server based on the numerical changes in the fault prediction coefficient sequence, and is specifically used to: determine that the failure risk of the server is increased when the numerical value of the fault prediction coefficient sequence shows an increasing trend; and determine that the failure risk of the server is reduced when the numerical value of the fault prediction coefficient sequence shows a decreasing trend and stops changing when reaching a preset value.

[0138] It should be noted that for details not disclosed in the server fault diagnosis device of the embodiment of the present application, please refer to the details disclosed in the server fault diagnosis method of the above embodiment of the present application, and the details will not be repeated here.

[0139] According to the server fault diagnosis device of the embodiment of the present application, the acquisition module acquires data information of multiple characteristic indicators of the server and the preset normal threshold range of each characteristic indicator, the first determination module determines the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range, the second determination module determines the characteristic comprehensive evaluation value according to the characteristic deviation of multiple characteristic indicators and their corresponding weight coefficients, and the fault diagnosis module performs fault diagnosis on the server according to the characteristic comprehensive evaluation value. Thus, the device performs characteristic diagnosis on the server through multiple characteristic indicators, improves the fault diagnosis accuracy, and ensures the operation stability of the server.

[0140] Corresponding to the above embodiments, the present application also proposes a computer-readable storage medium.

[0141] The computer-readable storage medium of the embodiment of the present application stores a server fault diagnosis program thereon, and the server fault diagnosis program implements the above-mentioned server fault diagnosis method when executed by a processor.

[0142] According to the computer-readable storage medium of the embodiment of the present application, the above-mentioned server fault diagnosis method is implemented when the server fault diagnosis program stored thereon is executed by the processor. Based on the above-mentioned server fault diagnosis method, feature diagnosis of the server is performed through multiple feature indicators, which improves the fault diagnosis accuracy and ensures the operation stability of the server.

[0143] Corresponding to the above embodiment, the present application also proposes an electronic device.

[0144] like Figure 5 As shown, the electronic device 100 of an embodiment of the present application includes a memory 110, a processor 120, and a server fault diagnosis program stored in the memory 110 and executable on the processor 120. When the processor 120 executes the server fault diagnosis program, the above-mentioned server fault diagnosis method is implemented.

[0145] According to the electronic device of the embodiment of the present application, when the server fault diagnosis program is executed by the processor, the above-mentioned server fault diagnosis method is implemented. Based on the above-mentioned server fault diagnosis method, feature diagnosis of the server is performed through multiple feature indicators, which improves the fault diagnosis accuracy and ensures the operation stability of the server.

[0146] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0147] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0148] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0149] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0150] In this application, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0151] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A server fault diagnosis method, characterized in that: The method comprises: Acquire data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator; Determine the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range; Determine a comprehensive feature evaluation value based on the feature deviations of the plurality of feature indicators and their corresponding weight coefficients; Fault diagnosis is performed on the server according to the comprehensive evaluation value of the features.

2. The server fault diagnosis method according to claim 1, characterized in that: Determining the characteristic deviation of the corresponding characteristic indicator according to the data information of each characteristic indicator and the corresponding preset normal threshold range includes: When the data information of the characteristic indicator is within the preset normal threshold range, determining that the characteristic deviation of the corresponding characteristic indicator is zero; When the data information of the characteristic indicator exceeds the upper limit of the corresponding preset normal threshold range, a first difference between the data information and the upper limit of the preset normal threshold range and a second difference between the upper limit and the lower limit of the preset normal threshold range are respectively obtained, and a characteristic deviation of the corresponding characteristic indicator is determined according to a ratio of the first difference to the second difference; When the data information of the characteristic indicator is lower than the lower limit value of the preset normal threshold range, the third difference between the lower limit value of the preset normal threshold range and the data information and the second difference between the upper limit value and the lower limit value of the preset normal range are obtained respectively, and the characteristic deviation of the corresponding characteristic indicator is determined according to the ratio of the third difference to the second difference.

3. The server fault diagnosis method according to claim 1, characterized in that: Determining the feature comprehensive evaluation value according to the feature deviations of the plurality of feature indicators and their corresponding weight coefficients includes: Obtaining a fault assessment value of a corresponding characteristic indicator according to the product of the characteristic deviation degree of the characteristic indicator and its corresponding weight coefficient; The feature comprehensive evaluation value is obtained according to the sum of the fault evaluation values ​​of the plurality of feature indicators.

4. The server fault diagnosis method according to claim 1, characterized in that: The performing fault diagnosis on the server according to the feature comprehensive evaluation value comprises: When the comprehensive evaluation value of the feature is greater than or equal to a preset fault threshold, determining that the server is in a high-risk state of failure; When the comprehensive evaluation value of the feature is less than a preset fault threshold, it is determined that the server is in a normal state or a low-fault risk state.

5. The server fault diagnosis method according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtain comprehensive evaluation values ​​of multiple historical features within a preset time period; Determining a fault prediction coefficient sequence within the preset time period according to the multiple historical feature comprehensive evaluation values; Fault prediction is performed on the server according to the numerical changes of the fault prediction coefficient sequence.

6. The server fault diagnosis method according to claim 5, characterized in that: The fault prediction coefficient sequence includes multiple fault prediction coefficients, and the fault prediction coefficients in the fault prediction coefficient sequence within the preset time period are calculated by the following formula: Among them, P k represents the kth fault prediction coefficient in the fault prediction coefficient sequence, E t represents the t-th historical feature comprehensive evaluation value within the preset time period, The average value of the comprehensive evaluation values ​​of multiple historical features within the preset time period, E t-k It represents the comprehensive evaluation value of the tkth historical feature within the preset time period.

7. The server fault diagnosis method according to claim 5, characterized in that: The performing fault prediction on the server according to the numerical change of the fault prediction coefficient sequence includes: In a case where the values ​​of the failure prediction coefficient sequence show an increasing trend, determining that the failure risk of the server is increasing; When the values ​​of the fault prediction coefficient sequence show a decreasing trend and stop changing when reaching a preset value, it is determined that the failure risk of the server is reduced.

8. A server fault diagnosis device, characterized in that: The device comprises: An acquisition module, used to acquire data information of multiple characteristic indicators of the server and a preset normal threshold range of each characteristic indicator; A first determination module is used to determine the characteristic deviation of the corresponding characteristic indicator according to the data information of the characteristic indicator and the corresponding preset normal threshold range; A second determination module is used to determine a feature comprehensive evaluation value according to the feature deviations of the plurality of feature indicators and their corresponding weight coefficients; A fault diagnosis module is used to perform fault diagnosis on the server according to the comprehensive evaluation value of the characteristics.

9. A computer-readable storage medium, characterized in that: A server fault diagnosis program is stored thereon, and when the server fault diagnosis program is executed by a processor, the server fault diagnosis method according to any one of claims 1-7 is implemented.

10. An electronic device, characterized in that: The invention comprises a memory, a processor and a server fault diagnosis program stored in the memory and executable on the processor. When the processor executes the server fault diagnosis program, the server fault diagnosis method according to any one of claims 1 to 7 is implemented.

Images