Software database exception monitoring and processing method

Abstract

The invention relates to the technical field of electric digital data processing, in particular to a software database exception monitoring and processing method which comprises the following steps: monitoring multi-dimensional performance indexes in database operation in real time, adopting a model to predict a future database state, dynamically adjusting weights through an error feedback mechanism, and performing data processing. Capturing state changes of the database in different time periods; and on the basis of a prediction result, identifying anomalies, constructing an abnormal state matrix, judging whether the database is in an abnormal state or not, starting a dynamic response mechanism, executing differentiated processing for different anomalies, allocating resources through a dynamic response function, preferentially processing serious anomalies, and optimizing a future processing strategy by analyzing historical abnormal processing data. According to the method, the monitoring sensitivity can be adaptively adjusted, and potential abnormal behaviors can be timely and effectively captured; and through a dynamic response mechanism, resources are intelligently allocated according to the abnormities of different types and severity degrees, so that the stability of the system is improved.

Description

Technical Field

[0001] The invention relates to the technical field of electronic digital data processing, and in particular to a software database abnormality monitoring and processing method. Background Art

[0002] In the modern information society, databases, as the core data storage and processing infrastructure, are widely used in various software systems and support the daily operation of various industries. With the rapid growth of data volume and the complexity of business needs, database systems not only bear the storage needs of massive data, but also need to support high-concurrency queries, transaction processing, and complex business logic operations. Therefore, the stability and efficient operation of the database are directly related to the reliability and performance of the entire system.

[0003] Although database technology has made significant progress in recent years, database systems still face various potential risks and challenges during operation. With the increase in user access and the complexity of business scenarios, the load pressure of the database has risen sharply, and various abnormal situations (such as query delays, transaction deadlocks, resource contention, data corruption, connection interruptions, etc.) have frequently occurred. If these abnormalities cannot be handled promptly and effectively, they will not only affect the performance of the system and lead to a decline in user experience, but may also cause serious data loss or service interruption, thus affecting the normal operation of the entire enterprise.

[0004] However, the existing technology is unable to adaptively adjust the sensitivity of monitoring according to the actual load conditions, resulting in frequent false alarms and missed alarms under high load, and unable to capture potential abnormal behaviors in a timely and effective manner. The existing technology lacks a dynamic response mechanism and cannot intelligently allocate resources according to abnormalities of different types and severity. It is easy for low-priority issues to be processed prematurely while high-priority issues are delayed, thereby increasing the instability and unreliability of the system. Summary of the invention

[0005] In order to solve the problem that the monitoring sensitivity cannot be adaptively adjusted according to the actual load conditions, resulting in frequent false alarms and missed alarms under high loads and the inability to capture potential abnormal behaviors in a timely and effective manner, the present invention provides a software database anomaly monitoring and processing method.

[0006] The present invention provides a software database abnormality monitoring and processing method, which adopts the following technical solutions: A software database abnormality monitoring and processing method comprises the following steps: Real-time monitoring of multi-dimensional performance indicators of database operation, using models to predict future database status, dynamically adjusting weights to optimize the model through error feedback mechanisms, and capturing database status changes in different time periods; Identify anomalies based on prediction results, build an abnormal status matrix, determine whether the database is in an abnormal state, and start a dynamic response mechanism to perform differentiated processing for different anomalies. Allocate resources through dynamic response functions, give priority to serious anomalies, and optimize future processing strategies by analyzing historical exception processing data.

[0007] In a specific feasible implementation scheme, the multi-dimensional performance indicators include query execution time, number of database connections, CPU and memory resource usage, transaction processing time, and disk I / O performance.

[0008] In a specific implementation scheme, in using a model to predict the future database state, a recursive piecewise weighted regression model is used, and the following steps are also included: The index values ​​of the collected multi-dimensional performance indicators constitute an initial state matrix; The initial state matrix is ​​divided into several time periods, and the state information in each time period is expressed by the matrix express:

[0009] in, Indicates the number of multi-dimensional performance indicators; Indicates the moment; The multi-dimensional performance indicators of the database are segmented and weighted to capture complex time-varying trends, dynamically estimate the database status at the current moment and in the future, and recursively predict future database status changes in each time period.

[0010] In a specific feasible implementation scheme, the recursive formula is:

[0011] In the above formula, for Moment Recursive estimates of multi-dimensional performance indicators; It is Within a period of time The weight of the recursive estimate at each moment; It is a coefficient used to adjust the recursive speed, adjusting the current time for the The impact of multiple dimensional performance indicators; Controls the time recursion speed.

[0012] In a specific implementation scheme, the weights are dynamically adjusted to optimize the model through an error feedback mechanism, and the state changes of the database in different time periods are captured, including the following steps: After each recursive prediction value is calculated, it is compared with the actual observed value to calculate the prediction error between the two. When calculating the prediction error, the smoothing coefficient is introduced in combination with the amplitude of relative change; The calculation formula for the prediction error is:

[0013] In the above formula, It is In the time period The prediction error of a multi-dimensional performance indicator; It is The number of moments in a time period; No. A collection of moments within a time period; yes Moment Actual observed values ​​of multi-dimensional performance indicators; is the error smoothing parameter.

[0014] The weight of the recursive piecewise weighted regression model is dynamically adjusted according to the prediction error of each time period. The weight of the time period with larger error will be reduced in subsequent calculations, and the weight of the time period with smaller error will be increased accordingly. When adjusting the weight, a sinusoidal function adjustment factor is introduced to periodically adjust the weight in the time dimension. The adjustment formula is:

[0015] in, is the adjustment amplitude of time period weight.

[0016] In a specific implementation scheme, identifying anomalies based on prediction results, constructing an abnormal state matrix, and determining whether the database is in an abnormal state include the following steps: Calculate the outlier value based on the recursive prediction result after weighted adjustment and generate an abnormal state matrix. If the outlier value of a multi-dimensional performance indicator continues to exceed the threshold, it is considered that the current multi-dimensional performance indicator is abnormal. The construction formula of the abnormal state matrix is:

[0017] in, Indicates Multi-dimensional performance indicators Outliers at the moment.

[0018] In a specific feasible implementation plan, starting the dynamic response mechanism, performing differentiated processing for different exceptions, allocating resources through the dynamic response function, and giving priority to processing serious exceptions include the following steps: Implement different processing strategies according to different exception types and the severity of the exceptions; The response function combines the degree of deviation of each abnormal indicator with the dynamic weight adjustment strategy to allocate resources according to the severity of the abnormality; The response function is as follows:

[0019] in, yes The total response value of the dynamic response mechanism to anomalies at any moment; No. The response weight of the multi-dimensional performance indicators is used to measure the priority of each performance indicator in exception handling; Indicates the relationship between abnormal deviation values ​​and normal values; is the response smoothing factor, which controls the smoothness of the response; It is a logic function used to control the abnormal response speed; Control the speed of abnormal response; For the The response threshold of a multi-dimensional performance indicator.

[0020] In a specific feasible implementation scheme, after executing the dynamic response mechanism, the response weight in the subsequent recursive process is adjusted according to the response result, and the adjustment formula is: .

[0021] In a specific implementation plan, by analyzing historical exception processing data, optimizing future processing strategies includes the following steps: Get historical exception processing data from the database, and construct the historical exception processing results and processing time into a historical data matrix:

[0022] in, is the historical data matrix; Indicates The treatment effect of the secondary anomaly; is the total number of exceptions handled; Through the recursive regression optimization model, historical data is analyzed to predict possible future processing needs and generate optimization regression results:

[0023] in, yes The optimized regression results at each moment; It is The treatment effect of the secondary anomaly; For the Sensitivity parameters of treatment effects of sub-anomalies; is the weight factor of historical data; The expected recovery time.

[0024] In a specific feasible implementation scheme, according to the optimized regression results, the subsequent processing strategy and sensitivity parameters are dynamically adjusted through the following adaptive adjustment formula:

[0025] in, is the time window to be considered.

[0026] In summary, the present invention has the following beneficial effects: 1. By dividing the time periods and weighting the performance indicators, short-term fluctuations can be captured, while the accuracy of long-term trend predictions is improved. By recursively calculating the future database state, the prediction capability of database anomaly detection is gradually optimized, so that abnormal behaviors can be better identified and analyzed, reducing the potential risks of database anomalies.

[0027] 2. The error feedback mechanism enables adaptive adjustment of the weight of recursive calculations during the forecasting process, and ensures more accurate future forecasts by continuously correcting the recursive forecast results. This feedback mechanism combines absolute error and relative change amplitude, enabling the system to respond quickly to short-term abnormal fluctuations and maintain high accuracy in long-term forecasts, further improving the reliability and adaptability of the system.

[0028] 3. After detecting an anomaly, automatically allocate processing resources and implement corresponding strategies according to different types of anomalies and the severity of the anomaly. The response mechanism can differentiate the priorities of different anomalies to ensure that high-priority anomalies are responded to and resolved faster, thereby improving the efficiency of handling anomalies and avoiding greater damage caused by the spread of anomalies.

[0029] 4. Combined with the recursive regression optimization model of historical processing data, the future processing strategy is continuously optimized by learning historical exception processing data. By recursively analyzing historical processing data, it is possible to effectively predict possible exception processing needs in the future, optimize exception processing strategies, and gradually improve processing efficiency. This adaptive optimization based on historical data improves the degree of intelligence and can self-improve and optimize the processing process based on the accumulated historical experience, thereby improving overall performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of the software database abnormality monitoring and processing method. DETAILED DESCRIPTION

[0031] The following combination Figure 1The present invention is described in further detail.

[0032] Reference Figure 1 ,The software database abnormality monitoring and processing method includes the following steps: S1,real-time monitoring of multi-dimensional performance indicators in database operation,adopts recursive piecewise weighted regression model to predict future database status,and dynamically adjusts weights to optimize the model through error feedback mechanism,capturing the status changes of database in different time periods.

[0033] Multi-dimensional performance indicators include query execution time, number of database connections, CPU and memory resource usage, transaction processing time, disk I / O performance, etc. The collection of multi-dimensional performance indicators is achieved by directly reading database system logs, query execution plans, and system performance monitoring interfaces, and the collected indicator values ​​constitute the initial state matrix.

[0034] In the recursive piecewise weighted regression model, the database state information of each time period is organized into a matrix, and the recursive calculation is performed based on the piecewise matrix. That is, the initial state matrix is ​​divided into several time periods, and the state information in each time period is represented by the matrix express:

[0035] in, Represents the number of multi-dimensional performance indicators. By weighting the multi-dimensional performance indicators of the database in sections, complex time-varying trends are captured, so that the database status at the current moment and in the future can be dynamically estimated. By using the section-wise recursive method, the accuracy of the overall prediction can be improved while capturing short-term fluctuations. The recursive formula is as follows:

[0036] In the above formula, for Moment Recursive estimates of multi-dimensional performance indicators; It is Within a period of time The weight of the recursive estimate at each moment; It is a coefficient used to adjust the recursive speed, adjusting the current time for the The impact of multiple dimensional performance indicators; Control the time recursion speed. Recursively predict the state changes of the future database in each time period, thus providing a basis for subsequent abnormal identification and processing.

[0037] The error feedback mechanism is introduced to correct the results of recursive prediction. After each recursive prediction value is calculated, it will be compared with the actual observation value to calculate the prediction error between the two. The prediction error includes not only the direct absolute error, but also the amplitude of relative changes. The smoothing coefficient is introduced to avoid the error being infinitely amplified due to short-term fluctuations. Based on the size of the prediction error, the weights of the subsequent recursive piecewise weighted regression model are adaptively adjusted, the recursive calculation is corrected, and the recursive strategy is adjusted in time to ensure that the prediction model can be gradually optimized, thereby gradually reducing future prediction errors.

[0038] Specifically, the calculation formula of the prediction error is:

[0039] In the above formula, It is In the time period The prediction error of a multi-dimensional performance indicator represents the difference between the actual observed value and the recursive estimated value; It is The number of moments in a time period; No. A collection of moments within a time period; yes Moment The actual observed values ​​of multi-dimensional performance indicators; is the error smoothing parameter. By calculating the prediction error, the accuracy of the recursive prediction is monitored in real time to ensure that the recursive prediction process is adjusted in different time periods and the recursive results are further optimized.

[0040] Furthermore, the weights of the recursive piecewise weighted regression model are dynamically adjusted according to the prediction error of each time period, so as to capture the fluctuation characteristics of different time periods in future predictions. The time period with larger errors will have its corresponding weight reduced in subsequent calculations, and the weight of the time period with smaller errors will be increased accordingly, gradually enhancing the ability to adapt to the database status. At the same time, in order to avoid excessively drastic weight adjustments, a sinusoidal function adjustment factor is introduced to periodically adjust the weights in the time dimension, capturing the periodic changes in database performance in different time periods, thereby further improving the accuracy of the prediction. The specific adjustment formula is:

[0041] in, It is the adjustment amplitude of the time period weight, which controls the size of the weighted update. The weight is adjusted periodically using the sine function to capture the fluctuation characteristics of database performance in different time periods. The weight at the next moment is used to improve the adaptability of the recursive piecewise weighted model. After weighted recursive adjustment, the state changes of the database at different stages are captured and dynamically corrected according to the actual error.

[0042] S2, after identifying anomalies based on recursive prediction results, constructs an abnormal state matrix to determine whether the database is in an abnormal state, and starts a dynamic response mechanism to perform differentiated processing for different anomalies. It allocates resources through dynamic response functions, prioritizes serious anomalies, and analyzes historical anomaly processing data through a recursive regression optimization model to optimize future processing strategies and gradually improve the efficiency and response speed of exception processing.

[0043] Specifically, the outlier value is calculated based on the recursive prediction result after weighted adjustment, and the abnormal state matrix is ​​generated. The data in the abnormal state matrix is ​​used to determine whether the current database is in an abnormal state: if the outlier value of the multi-dimensional performance indicator continues to exceed the set threshold, it is considered that the current multi-dimensional performance indicator is abnormal. The outlier value is the difference between the recursive prediction result after weighted adjustment and the actual observation value.

[0044] The construction formula of the abnormal state matrix is ​​as follows:

[0045] in, Indicates Multi-dimensional performance indicators Once the abnormal value of one or more multi-dimensional performance indicators exceeds the set threshold, the dynamic response mechanism will be automatically triggered.

[0046] In the dynamic response mechanism, different processing strategies are executed according to different abnormal types and the severity of the abnormalities; the response function combines the deviation degree of each abnormal indicator with the dynamic weight adjustment strategy to allocate resources according to the severity of the abnormality. The response function of the dynamic response mechanism is as follows:

[0047] in, yes The total response value of the dynamic response mechanism to anomalies at any moment; No. The response weight of the multi-dimensional performance indicators is used to measure the priority of each performance indicator in exception handling; Indicates the relationship between abnormal deviation values ​​and normal values; is the response smoothing factor, which controls the smoothness of the response; It is a logic function used to control the abnormal response speed; Control the speed of abnormal response; For the The response function can be used to handle different types of exceptions differently, ensuring that exceptions with higher priority are responded to quickly.

[0048] After executing the dynamic response mechanism, the response weights in the subsequent recursive process are adjusted according to the response results to ensure that the response strategy can be optimized when dealing with similar exceptions. The adjustment formula is:

[0049] Adjusting the response weight of each indicator according to the severity of the anomaly can allocate more resources and processing capabilities to abnormal high-risk indicators in future recursive processes, thereby enhancing the anomaly handling effect.

[0050] When the exception processing is completed, the historical exception processing data is obtained from the database. By learning the historical exception processing data, the future processing strategy is continuously optimized, and the historical exception processing results and processing time are constructed into a historical data matrix:

[0051] in, is the historical data matrix; Indicates The processing effect of the secondary exception, including response time, recovery time and other data; is the total number of exception handling. Through the recursive regression optimization model, historical data is analyzed to predict possible future processing requirements and generate optimization regression results:

[0052] in, yes The optimized regression result at each moment is an evaluation of the effect of historical anomaly handling; It is The treatment effect of the secondary anomaly; For the The sensitivity parameter of the treatment effect of the secondary anomaly controls the impact of the treatment effect on the optimization result; is the weight factor of historical data, which controls the influence weight of historical processing data on the current optimization result; is the expected recovery time, which represents the processing time under ideal conditions. Through recursive analysis of historical processing data, the exception handling strategy is gradually optimized to improve the efficiency and response speed of future processing.

[0053] According to the optimized regression results, the following adaptive adjustment formula is used to dynamically adjust the subsequent processing strategy and sensitivity parameters to ensure the efficiency of exception handling is gradually improved:

[0054] in, It is the time window considered, which represents the cumulative value of all historical optimization results during this period. Based on long-term historical data, more robust decision adjustments can be made. By optimizing the regression results, the processing sensitivity is adjusted to ensure that future processing strategies can be adaptively optimized based on historical processing experience.

[0055] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A software database abnormality monitoring and processing method, characterized in that: The steps include: Real-time monitoring of multi-dimensional performance indicators of database operation, using models to predict future database status, dynamically adjusting weights to optimize the model through error feedback mechanisms, and capturing database status changes in different time periods; Identify anomalies based on prediction results, build an abnormal status matrix, determine whether the database is in an abnormal state, and start a dynamic response mechanism to perform differentiated processing for different anomalies. Allocate resources through dynamic response functions, give priority to serious anomalies, and optimize future processing strategies by analyzing historical exception processing data.

2. The software database abnormality monitoring and processing method according to claim 1, characterized in that: Multi-dimensional performance indicators include query execution time, number of database connections, CPU and memory resource usage, transaction processing time, and disk I / O performance.

3. The software database abnormality monitoring and processing method according to claim 1, characterized in that: In the process of using a model to predict the future database state, a recursive piecewise weighted regression model is used, which also includes the following steps: The index values ​​of the collected multi-dimensional performance indicators constitute an initial state matrix; The initial state matrix is ​​divided into several time periods, and the state information in each time period is expressed by the matrix express: in, Indicates the number of multi-dimensional performance indicators; Indicates the moment; The multi-dimensional performance indicators of the database are segmented and weighted to capture complex time-varying trends, dynamically estimate the database status at the current moment and in the future, and recursively predict future database status changes in each time period.

4. The software database abnormality monitoring and processing method according to claim 3 is characterized in that: The recursive formula is: In the above formula, for Moment Recursive estimates of multi-dimensional performance indicators; It is Within a period of time The weight of the recursive estimate at each moment; It is a coefficient used to adjust the recursive speed, adjusting the current time for the The impact of multiple dimensional performance indicators; Controls the time recursion speed.

5. The software database abnormality monitoring and processing method according to claim 4 is characterized in that: Through the error feedback mechanism, the weights are dynamically adjusted to optimize the model. Capturing the state changes of the database in different time periods includes the following steps: After each recursive prediction value is calculated, it is compared with the actual observed value to calculate the prediction error between the two. When calculating the prediction error, the smoothing coefficient is introduced in combination with the amplitude of relative change; The calculation formula for the prediction error is: In the above formula, It is In the time period The prediction error of a multi-dimensional performance indicator; It is The number of moments in a time period; No. A collection of moments within a time period; yes Moment Actual observed values ​​of multi-dimensional performance indicators; is the error smoothing parameter; The weight of the recursive piecewise weighted regression model is dynamically adjusted according to the prediction error of each time period. The weight of the time period with larger error will be reduced in subsequent calculations, and the weight of the time period with smaller error will be increased accordingly. When adjusting the weight, a sinusoidal function adjustment factor is introduced to periodically adjust the weight in the time dimension. The adjustment formula is: in, is the adjustment amplitude of time period weight.

6. The software database abnormality monitoring and processing method according to claim 1, characterized in that: Identifying anomalies based on prediction results, building an abnormal state matrix, and determining whether the database is in an abnormal state include the following steps: Calculate the outlier value based on the recursive prediction result after weighted adjustment and generate an abnormal state matrix. If the outlier value of a multi-dimensional performance indicator continues to exceed the threshold, it is considered that the current multi-dimensional performance indicator is abnormal. The construction formula of the abnormal state matrix is: in, Indicates Multi-dimensional performance indicators Outliers at the moment.

7. The software database abnormality monitoring and processing method according to claim 6 is characterized in that: Starting the dynamic response mechanism, performing differentiated processing for different exceptions, allocating resources through dynamic response functions, and giving priority to handling serious exceptions include the following steps: Implement different processing strategies according to different exception types and the severity of the exceptions; The response function combines the degree of deviation of each abnormal indicator with the dynamic weight adjustment strategy to allocate resources according to the severity of the abnormality; The response function is as follows: in, yes The total response value of the dynamic response mechanism to anomalies at any moment; No. The response weight of the multi-dimensional performance indicators is used to measure the priority of each performance indicator in exception handling; Indicates the relationship between abnormal deviation values ​​and normal values; is the response smoothing factor, which controls the smoothness of the response; It is a logic function used to control the abnormal response speed; Control the speed of abnormal response; For the The response threshold of a multi-dimensional performance indicator.

8. The software database abnormality monitoring and processing method according to claim 7 is characterized in that: After executing the dynamic response mechanism, the response weight in the subsequent recursive process is adjusted according to the response result. The adjustment formula is: 。 9. The software database abnormality monitoring and processing method according to claim 1, characterized in that: By analyzing historical exception handling data, optimizing future handling strategies includes the following steps: Get historical exception processing data from the database, and construct the historical exception processing results and processing time into a historical data matrix: in, is the historical data matrix; Indicates The treatment effect of the secondary anomaly; is the total number of exceptions handled; Through the recursive regression optimization model, historical data is analyzed to predict possible future processing needs and generate optimization regression results: in, yes The optimized regression results at each moment; It is The treatment effect of the secondary anomaly; For the Sensitivity parameters of treatment effects of sub-anomalies; is the weight factor of historical data; The expected recovery time.

10. The software database abnormality monitoring and processing method according to claim 9, characterized in that: According to the optimized regression results, the subsequent processing strategies and sensitivity parameters are dynamically adjusted through the following adaptive adjustment formula: in, is the time window to be considered.

Images