A storage system performance indicator abnormality detection method, device and storage medium

Abstract

The present invention discloses a storage system performance indicator anomaly detection method, device and storage medium, which relates to the field of storage technology, including: obtaining a first performance indicator time series data set corresponding to a performance indicator to be detected in a storage system; respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set; segmenting the first performance indicator time series data set according to each difference to obtain each performance indicator subset; calculating the performance indicator threshold corresponding to each performance indicator subset; and performing performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold. The storage system performance indicator anomaly detection method provided by the present invention can provide a more accurate and more adaptable dynamic threshold to data characteristics, thereby improving the accuracy of anomaly detection.

Description

Technical Field

[0001] The present invention relates to the field of storage technology, and in particular to a method, device, computer-readable storage medium and computer program product for detecting abnormality of storage system performance indicators. Background Art

[0002] The operation status and performance level of the storage system will affect the system's data processing efficiency and system stability. As the complexity of storage systems continues to increase, storage system performance anomalies have become an important factor affecting system stability and user experience.

[0003] Storage system performance anomalies are reflected in the anomalies of performance indicators. Performance indicator anomalies are usually manifested as performance indicators exceeding their set thresholds at a certain moment. Therefore, the core task of indicator anomaly detection is to detect whether storage system performance indicators exceed the thresholds. In practical applications, the establishment of storage system performance indicator thresholds is an important issue. Setting the threshold too high may result in the failure to detect performance anomalies in a timely manner, while setting the threshold too low may cause too many false alarms. Improper threshold setting may lead to inaccurate storage system performance indicator anomaly detection results.

[0004] In summary, how to effectively solve the problem of inaccurate storage system performance indicator anomaly detection results caused by inappropriate threshold settings is a problem that technicians in this field urgently need to solve. Summary of the invention

[0005] An object of the present invention is to provide a method for detecting anomalies in performance indicators of a storage system, which method can provide a more accurate dynamic threshold that is more adaptable to data characteristics, thereby improving the accuracy of anomaly detection; another object of the present invention is to provide a storage system performance indicator anomaly detection device, a computer-readable storage medium, and a computer program product.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A method for detecting abnormality of storage system performance indicators, comprising:

[0008] Acquire a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system;

[0009] Respectively calculating the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set;

[0010] Segmenting the first performance indicator time series data set according to each difference value to obtain each performance indicator subset;

[0011] Calculate the performance indicator thresholds corresponding to each performance indicator subset;

[0012] According to each performance indicator threshold, performance indicator anomaly detection is performed on each performance indicator subset.

[0013] In a specific implementation of the present invention, after obtaining a first performance indicator time series data set corresponding to a performance indicator to be detected in a storage system, and before respectively calculating the difference between first performance indicators of adjacent time series in the first performance indicator time series data set, the method further includes:

[0014] Normalization is performed on each first performance indicator in the first performance indicator time series data set.

[0015] In a specific implementation of the present invention, segmenting the first performance indicator time series data set according to each difference value includes:

[0016] Sort the differences by size to obtain a difference sequence;

[0017] Filtering a preset number of target differences from the end with the larger difference in the difference sequence;

[0018] Obtain target performance indicator pairs of adjacent time series corresponding to each target difference respectively;

[0019] The first performance indicator time series data set is segmented by taking the middle position of the first performance indicator of the previous time series and the first performance indicator of the next time series contained in each target performance indicator pair as the segmentation point.

[0020] In a specific implementation of the present invention, calculating the performance indicator thresholds corresponding to each performance indicator subset includes:

[0021] Calculate the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset respectively;

[0022] The difference between the mean of each performance indicator and three times the standard deviation of the corresponding performance indicator is calculated to obtain the lower limit of the performance indicator corresponding to each performance indicator subset;

[0023] The mean of each performance indicator and three times the standard deviation of the corresponding performance indicator are summed up to obtain the upper limit of the performance indicator corresponding to each performance indicator subset;

[0024] The performance indicator thresholds corresponding to each performance indicator subset are determined according to the performance indicator lower limits and performance indicator upper limits corresponding to each performance indicator subset.

[0025] In a specific implementation of the present invention, after obtaining each performance indicator subset, the method further includes:

[0026] Obtain the maximum performance index in each performance index subset respectively;

[0027] Each performance indicator subset is recorded using the performance indicator mean and the maximum performance indicator respectively.

[0028] In a specific implementation of the present invention, the performance indicator anomaly detection is performed on each performance indicator subset according to each performance indicator threshold, including:

[0029] Compare the first performance indicator in each performance indicator subset with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold respectively to obtain a comparison result;

[0030] When the comparison result is less than a lower limit of the performance indicator or greater than an upper limit of the performance indicator, it is determined that the first performance indicator is an abnormal performance indicator.

[0031] In a specific implementation of the present invention, after determining that the first performance indicator is an abnormal performance indicator, the method further includes:

[0032] Determine whether there are multiple consecutive abnormal performance indicators;

[0033] If so, determine the target time range corresponding to multiple consecutive abnormal performance indicators;

[0034] The starting position and the ending position of the abnormal interval composed of a plurality of continuous abnormal performance indicators are determined according to the target time range.

[0035] In a specific implementation of the present invention, after determining the starting position and the ending position of the abnormal interval composed of a plurality of continuous abnormal performance indicators according to the target time range, the method further includes:

[0036] Selecting a first indicator sequence to be detected for correlation from the first performance indicator time series data set according to the abnormal interval;

[0037] Obtain a second performance indicator time series data set to be tested for correlation;

[0038] Filtering each second performance indicator corresponding to the time range to which the first indicator series belongs from the second performance indicator time series data set;

[0039] Calculating the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators;

[0040] The correlation between the first performance indicator and the second performance indicator is determined according to the correlation coefficient.

[0041] In a specific implementation of the present invention, selecting the first indicator sequence to be detected for correlation from the first performance indicator time series data set according to the abnormal interval includes:

[0042] Calculating a first duration corresponding to the target time range;

[0043] Divide the first duration into two equal parts to obtain a second duration;

[0044] Filter the first performance indicators corresponding to the first second duration of the target time range from the first performance indicator time series data set, and filter the first performance indicators corresponding to the second duration of the target time range;

[0045] The first performance indicators obtained by screening are determined as the first indicator sequence.

[0046] In a specific implementation of the present invention, determining the correlation between the first performance indicator and the second performance indicator according to the correlation coefficient includes:

[0047] When the absolute value of the difference between the correlation coefficient and 1 is less than a first preset value, it is determined that the first performance indicator is positively correlated with the second performance indicator;

[0048] When the absolute value of the difference between the correlation coefficient and -1 is less than a second preset value, it is determined that the first performance indicator is negatively correlated with the second performance indicator;

[0049] When the absolute value of the difference between the correlation coefficient and 0 is smaller than a third preset value, it is determined that the first performance indicator is independent of the second performance indicator.

[0050] In a specific implementation of the present invention, obtaining a second performance indicator time series data set to be detected for correlation includes:

[0051] Determine from the storage system the object to be detected to which the first performance indicator time series data set belongs;

[0052] Filtering the target indicator type to be detected for correlation from the indicator type set corresponding to the object to be detected;

[0053] A performance indicator time series data set of a target indicator type corresponding to the object to be detected is determined as the second performance indicator time series data set.

[0054] In a specific embodiment of the present invention, after obtaining a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system, and before respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set, the method further includes:

[0055] When it is determined that there are missing values ​​in the first performance indicator time series data set, obtaining a first performance indicator corresponding to a previous time series of the target time series to which the missing value belongs;

[0056] Determine the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs as the performance indicator to be filled;

[0057] Fill the performance indicator to be filled into the corresponding position of the target time series in the first performance indicator time series data set.

[0058] A storage system performance indicator anomaly detection device, comprising:

[0059] Memory for storing computer programs;

[0060] The processor is used to implement the steps of the storage system performance indicator abnormality detection method as described above when executing the computer program.

[0061] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the storage system performance indicator anomaly detection method as described above are implemented.

[0062] A computer program product includes a computer program, wherein when the computer program is executed by a processor, the steps of the storage system performance indicator abnormality detection method as described above are implemented.

[0063] The storage system performance indicator anomaly detection method provided by the present invention obtains a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system; respectively calculates the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set; segments the first performance indicator time series data set according to each difference to obtain each performance indicator subset; calculates the performance indicator threshold value corresponding to each performance indicator subset; and performs performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold value.

[0064] The beneficial effect of the present invention is that by respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set, the first performance indicator time series data set is segmented according to each difference, and an adaptive segmentation processing strategy is adopted, which can dynamically adjust the segment length according to the characteristics of the data itself, which is more flexible than the fixed-length segmentation method, and can more accurately adapt to the local changes of the time series, thereby providing a more accurate and dynamic threshold that is more adapted to the data characteristics. By comparing each first performance indicator in each performance indicator subset obtained by segmentation with the calculated performance indicator threshold, outliers can be accurately identified, accurate analysis of storage system performance indicators is achieved, and the accuracy of anomaly detection is improved.

[0065] Correspondingly, the present invention also provides a storage system performance indicator anomaly detection device, a computer-readable storage medium and a computer program product corresponding to the above-mentioned storage system performance indicator anomaly detection method, which have the above-mentioned technical effects and are not repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1 A flowchart of an implementation of a method for detecting abnormal performance indicators of a storage system in an embodiment of the present invention;

[0068] Figure 2 Another implementation flow chart of the storage system performance indicator abnormality detection method in an embodiment of the present invention;

[0069] Figure 3 A schematic diagram of an adaptive threshold setting in an embodiment of the present invention;

[0070] Figure 4 This is a structural block diagram of a storage system performance indicator abnormality detection device in an embodiment of the present invention;

[0071] Figure 5 This is a structural block diagram of a storage system performance indicator anomaly detection device in an embodiment of the present invention;

[0072] Figure 6 A schematic diagram of the specific structure of a storage system performance indicator anomaly detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0073] In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0074] See also Figure 1 , Figure 1 This is a flowchart of an implementation of a method for detecting abnormality in storage system performance indicators in an embodiment of the present invention. The method may include the following steps:

[0075] S101: Acquire a first performance indicator time series data set corresponding to a performance indicator to be detected in a storage system.

[0076] The current performance indicator to be detected is determined in advance. The performance indicator to be detected may be the number of input / output operations per second (IOPS), bandwidth, latency, central processing unit (CPU) utilization, etc. A first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system is obtained. The first performance indicators in the first performance indicator time series data set are arranged in the order of generation time.

[0077] S102: Calculate the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set respectively.

[0078] After obtaining the first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system, the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set is calculated respectively. By calculating the difference between the first performance indicators of adjacent time series, it can be determined which two adjacent time series have a larger difference in first performance indicators.

[0079] S103: Segment the first performance indicator time series data set according to each difference value to obtain each performance indicator subset.

[0080] After calculating the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set, the first performance indicator time series data set is segmented according to the difference. For example, the positions where the difference fluctuations are relatively large can be determined as segmentation points for segmentation to obtain various performance indicator subsets.

[0081] S104: Calculate the performance indicator thresholds corresponding to each performance indicator subset.

[0082] After each performance indicator subset is obtained by segmentation, the performance indicator threshold corresponding to each performance indicator subset is calculated, so that the performance indicator threshold can be calculated according to the characteristics of each performance indicator subset itself, which greatly improves the accuracy of threshold setting compared to setting a unified fixed threshold.

[0083] S105: Performing performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold.

[0084] After calculating the performance indicator thresholds corresponding to each performance indicator subset, each performance indicator subset is subjected to performance indicator anomaly detection according to each performance indicator threshold, and the first performance indicator exceeding the threshold is determined as an abnormal indicator. This achieves the scientific and reasonable determination of the performance indicator thresholds and improves the accuracy of indicator anomaly detection.

[0085] It can be seen from the above technical solution that by respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set, the first performance indicator time series data set is segmented according to each difference, and an adaptive segmentation processing strategy is adopted. The segmentation length can be dynamically adjusted according to the characteristics of the data itself. Compared with the fixed-length segmentation method, it is more flexible and can more accurately adapt to the local changes of the time series, thereby providing a more accurate and dynamic threshold that is more adapted to the data characteristics. By comparing each first performance indicator in each performance indicator subset obtained by segmentation with the calculated performance indicator threshold, outliers can be accurately identified, accurate analysis of storage system performance indicators is achieved, and the accuracy of anomaly detection is improved.

[0086] It should be noted that, based on the above embodiment, the embodiment of the present invention also provides corresponding improved solutions. In the subsequent embodiments, the same steps or corresponding steps as those in the above embodiment can be referenced to each other, and the corresponding beneficial effects can also be referenced to each other, which will not be repeated one by one in the following improved embodiments.

[0087] See also Figure 2 , Figure 2 Another implementation flow chart of the storage system performance indicator abnormality detection method in an embodiment of the present invention, the method may include the following steps:

[0088] S201: Acquire a first performance indicator time series data set corresponding to a performance indicator to be detected in a storage system.

[0089] S202: When it is determined that there are missing values ​​in the first performance indicator time series data set, obtain a first performance indicator corresponding to a previous time series of the target time series to which the missing value belongs.

[0090] The acquired first performance indicator time series data set may contain missing first performance indicator time series. When it is determined that there are missing values ​​in the first performance indicator time series data set, the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs is obtained.

[0091] S203: Determine the first performance indicator corresponding to the previous time series of the target time series to which the missing value is obtained as the performance indicator to be filled.

[0092] After obtaining the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs, the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs is determined as the performance indicator to be filled. By determining the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs as the performance indicator to be filled, the filled indicator value and the indicator value of the previous time series are kept stable, avoiding the introduction of erroneous segmentation points caused by excessive fluctuations of the filled indicator value compared with the indicator value of the previous time series, and ensuring the continuity and integrity of the performance indicator.

[0093] S204: Fill the performance indicator to be filled into the corresponding position of the target time series in the first performance indicator time series data set.

[0094] After determining the performance indicator to be filled, the performance indicator to be filled is filled into the corresponding position of the target time series in the first performance indicator time series data set, thereby achieving effective filling of missing values.

[0095] S205: Normalizing each first performance indicator in the first performance indicator time series data set.

[0096] After the missing values ​​in the first performance indicator time series data set are filled, each first performance indicator in the first performance indicator time series data set is normalized. By normalizing each first performance indicator, each first performance indicator is scaled to a uniform range, thereby ensuring the accuracy and consistency of each first performance indicator.

[0097] When normalizing each first performance indicator, each coefficient can be divided by Normalize, where is the discrete Haar wavelet transform (Haar) resolution scale.

[0098] S206: Calculate the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set respectively.

[0099] S207: Sort the differences by size to obtain a difference sequence.

[0100] After calculating the difference between the first performance indicators of each adjacent time series, the difference values ​​are sorted to obtain a difference sequence. When sorting the difference values, the difference values ​​can be sorted in order from small to large, or in order from large to small.

[0101] S208: Filtering a preset number of target differences from the end of the difference sequence with a larger difference.

[0102] After the differences are sorted by size to obtain a difference sequence, a preset number of target differences are filtered from the end of the difference sequence with a larger difference, so as to filter out the larger difference.

[0103] It should be noted that the preset number can be set and adjusted according to actual conditions, and the embodiment of the present invention does not limit this.

[0104] S209: Obtain target performance indicator pairs of adjacent time series corresponding to each target difference value respectively.

[0105] After a preset number of target differences are screened out, target performance indicator pairs of adjacent time series corresponding to each target difference are respectively obtained, that is, target performance indicator pairs including first performance indicators of two adjacent time series are obtained.

[0106] S210: Using the middle position between the first performance indicator of the previous time series and the first performance indicator of the next time series included in each target performance indicator pair as a segmentation point, segment the first performance indicator time series data set to obtain each performance indicator subset.

[0107] After obtaining the target performance indicator pairs of adjacent time series corresponding to each target difference, the first performance indicator time series data set is segmented using the middle position of the first performance indicator of the previous time series and the first performance indicator of the next time series contained in each target performance indicator pair as the segmentation point to obtain each performance indicator subset, so that the first performance indicators of the two time series with larger waveforms are divided into two different performance indicator subsets, so that each first performance indicator in each performance indicator subset obtained by division is relatively stable. The length and number of segments are not pre-set, but are automatically determined based entirely on the characteristics of the data. In high-activity data areas, due to the rapid changes in data, shorter segments tend to be selected in order to more accurately capture the changing characteristics of performance indicators. On the contrary, in low-activity data areas, the performance indicators are relatively stable, so longer sequences are selected to divide data segments. Compared with the fixed-length segmentation method, the adaptive segmentation strategy proposed in the present invention does not rely on a pre-set segment length or number, but dynamically adjusts the segmentation according to the characteristics of the performance indicator itself, thereby dividing the first performance indicator time series data set into a series of performance indicator subsets with similar activities, which can more flexibly adapt to local changes in the time series, and thus provide a more accurate and more adaptive dynamic threshold to data characteristics.

[0108] When segmenting the first performance indicator time series data set, after respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set, sort the differences in chronological order to obtain a difference sequence, respectively calculate the fluctuation ratio of the difference corresponding to the latter time series in two adjacent differences in the difference sequence compared with the difference corresponding to the former time series; select the target difference whose waveform ratio exceeds the preset ratio from each difference; obtain the target performance indicator pair of the adjacent time series corresponding to the target difference, and use the middle position of the first performance indicator of the former time series and the first performance indicator of the latter time series contained in each target performance indicator pair as the segmentation point, segment the first performance indicator time series data set to obtain each performance indicator subset. By selecting the target difference with a large difference fluctuation ratio, the first performance indicator with abnormal fluctuation is accurately identified, and the segmentation accuracy is improved.

[0109] S211: Calculate the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset respectively.

[0110] After each performance indicator subset is obtained by division, the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset are calculated respectively.

[0111] S212: Calculate the difference between the mean of each performance indicator and three times the corresponding standard deviation of the performance indicator to obtain the lower limit of the performance indicator corresponding to each performance indicator subset.

[0112] After calculating the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset, the difference between each performance indicator mean and three times the corresponding performance indicator standard deviation is calculated to obtain the performance indicator lower limit corresponding to each performance indicator subset.

[0113] S213: Calculate the sum of the mean value of each performance indicator and three times the corresponding standard deviation of the performance indicator to obtain the upper limits of the performance indicators corresponding to each performance indicator subset.

[0114] After calculating the performance indicator mean and the performance indicator standard deviation corresponding to each performance indicator subset, the sum of each performance indicator mean and three times the corresponding performance indicator standard deviation is calculated to obtain the performance indicator upper limit corresponding to each performance indicator subset.

[0115] By taking the difference between the mean of each performance indicator and three times the corresponding standard deviation of the performance indicator as the lower limit of the performance indicator, and taking the sum of the mean of each performance indicator and three times the corresponding standard deviation of the performance indicator as the upper limit of the performance indicator, it can be ensured that the setting of the performance indicator threshold not only takes into account the normal fluctuations of the data, but also can effectively identify anomalies or extreme values.

[0116] S214: Determine the performance indicator thresholds corresponding to each performance indicator subset according to the performance indicator lower limit and the performance indicator upper limit corresponding to each performance indicator subset.

[0117] After calculating the performance indicator upper limit and the performance indicator lower limit corresponding to each performance indicator subset, the performance indicator threshold corresponding to each performance indicator subset is determined according to the performance indicator lower limit and the performance indicator upper limit corresponding to each performance indicator subset, so as to determine the performance indicator value between the performance indicator lower limit and the performance indicator upper limit corresponding to each performance indicator subset as the performance indicator value that meets the performance indicator threshold.

[0118] S215: Compare the first performance indicator in each performance indicator subset with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold respectively to obtain a comparison result.

[0119] After determining the performance indicator thresholds corresponding to each performance indicator subset, the first performance indicator in each performance indicator subset is compared with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold to obtain a comparison result, that is, to determine whether the first performance indicator in each performance indicator subset is a performance indicator value between the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold.

[0120] S216: When the comparison result is less than the lower limit of the performance indicator or greater than the upper limit of the performance indicator, determine that the first performance indicator is an abnormal performance indicator.

[0121] After obtaining the comparison result, when the comparison result is less than the lower limit of the performance indicator or greater than the upper limit of the performance indicator, it indicates that the first performance indicator exceeds the normal range, and the first performance indicator is determined to be an abnormal performance indicator. By comparing the first performance indicator in each performance indicator subset with the lower limit of the performance indicator and the upper limit of the performance indicator in the corresponding performance indicator threshold, rapid and accurate detection of abnormal performance indicators is achieved.

[0122] See also Figure 3 , Figure 3 The diagram is a schematic diagram of an adaptive threshold setting in an embodiment of the present invention. By comparing the first performance indicator in each performance indicator subset with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold, three locations exceeding the threshold can be obtained.

[0123] S217: Determine whether there are multiple consecutive abnormal performance indicators. If so, execute step S219; if not, do nothing.

[0124] After determining the abnormal performance indicator, determine whether there are multiple continuous abnormal performance indicators. If so, it means that there is an abnormal interval consisting of multiple continuous abnormal performance indicators, and execute step S218. If not, it means that the determined abnormal performance indicator is an accidental abnormality and no processing is performed.

[0125] S218: Determine a target time range corresponding to a plurality of continuous abnormal performance indicators.

[0126] When it is determined that there are multiple continuous abnormal performance indicators, it means that there is an abnormal interval composed of multiple continuous abnormal performance indicators, and the timing values ​​corresponding to each abnormal performance indicator are obtained, and the target time range corresponding to the multiple continuous abnormal performance indicators is determined according to each timing value.

[0127] S219: Determine the starting position and the ending position of the abnormal interval composed of multiple continuous abnormal performance indicators according to the target time range.

[0128] After determining the target time range corresponding to multiple consecutive abnormal performance indicators, the start and end positions of the abnormal intervals formed by multiple consecutive abnormal performance indicators are determined according to the target time range. By accurately identifying the abnormal intervals in which the performance indicators continuously exceed the performance indicator threshold, operation and maintenance personnel can quickly and accurately locate the root cause of the problem when facing performance abnormalities.

[0129] S220: Filtering a first indicator sequence to be detected for correlation from the first performance indicator time series data set according to the abnormal interval.

[0130] After determining the starting position and the ending position of the abnormal interval, it is indicated that the correlation between each first performance indicator and other performance indicators near the abnormal interval needs to be detected, and the first indicator sequence to be detected for correlation is selected from the first performance indicator time series data set according to the abnormal interval.

[0131] In a specific implementation of the present invention, step S220 may include the following steps:

[0132] Step 1: Calculate the first duration corresponding to the target time range;

[0133] Step 2: Divide the first duration into two equal parts to obtain the second duration;

[0134] Step 3: Filter the first performance indicators corresponding to the second time period before the target time range from the first performance indicator time series data set, and filter the first performance indicators corresponding to the second time period after the target time range;

[0135] Step 4: Determine the first performance indicators obtained by screening as a first indicator sequence.

[0136] For the convenience of description, the above four steps can be combined for explanation.

[0137] After determining the starting and ending positions of the abnormal interval, the first duration corresponding to the target time range is calculated, and the first duration is divided into two equal parts to obtain the second duration. The first performance indicators corresponding to the first second duration of the target time range are screened from the first performance indicator time series data set, and the first performance indicators corresponding to the second duration of the target time range are screened, and the first performance indicators obtained by screening are determined as the first indicator sequence. By using the first performance indicators of the second durations before and after the abnormal interval as the first indicator sequence for correlation analysis, the selected first indicator sequence contains sufficient context information, so as to more accurately capture and explain the correlation in the performance indicator time series. By comprehensively considering the lag difference to accurately measure the distance between the two time series, even if these changes are offset in time, the common change direction of the two performance indicators over time can be more effectively captured.

[0138] S221: Obtain a second performance indicator time series data set to be tested for correlation.

[0139] After the first indicator sequence to be detected for correlation is obtained through screening, a second performance indicator time series data set to be detected for correlation is obtained.

[0140] In a specific implementation of the present invention, step S221 may include the following steps:

[0141] Step 1: Determine the object to be detected to which the first performance indicator time series data set belongs from the storage system;

[0142] Step 2: Select the target indicator type to be detected from the indicator type set corresponding to the object to be detected;

[0143] Step three: Determine the performance indicator time series data set of the target indicator type corresponding to the object to be detected as the second performance indicator time series data set.

[0144] For the convenience of description, the above three steps can be combined for explanation.

[0145] The objects to be detected in the storage system may include systems, storage pools, volumes, hard disks, etc. The object to be detected to which the first performance indicator time series data set belongs is determined from the storage system, the target indicator type to be detected for correlation is screened from the indicator type set corresponding to the object to be detected, and the performance indicator time series data set of the target indicator type corresponding to the object to be detected is determined as the second performance indicator time series data set. By determining the object to be detected to which the first performance indicator time series data set belongs, and screening the target indicator type to be detected for correlation from the indicator type set corresponding to the object to be detected, correlation detection of different performance indicators of the same object is achieved.

[0146] In addition, the potential correlation analysis between indicators of different objects can be performed to help users quickly identify factors that may affect abnormal indicators, thereby improving the efficiency of problem location. By analyzing the mutual influence between different performance indicators of the same object, the correlation and dependency between them can be explored. By analyzing the potential correlation between different objects, possible cross-object influencing factors can be found.

[0147] S222: Filter the second performance indicators corresponding to the time range to which the first indicator series belongs from the second performance indicator time series data set.

[0148] After obtaining the second performance indicator time series data set to be detected for correlation, the second performance indicators corresponding to the time range to which the first indicator series belongs are screened from the second performance indicator time series data set, thereby screening out the second performance indicators belonging to the same time range as the first indicator series. By screening the performance indicators within the same time range, the screened second performance indicators used for correlation detection have more reference significance.

[0149] S223: Calculate the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators.

[0150] After the second performance indicators are screened, the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators is calculated.

[0151] Assume that the first indicator sequence and the second index sequence The common correlation coefficient It can be calculated by the following formula:

[0152] ;

[0153] ;

[0154] ;

[0155] ;

[0156] Where M represents the length of the time series, represents the value of the sequence x at time t, represents the value of the sequence y at time t, express The first-order autocorrelation coefficient of express The first-order autocorrelation coefficient of express and The mutual correlation coefficient.

[0157] S224: Determine the correlation between the first performance indicator and the second performance indicator according to the correlation coefficient.

[0158] After calculating the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators, the correlation between the first performance indicator and the second performance indicator is determined according to the correlation coefficient. Through correlation impact analysis, users can find the correlation factors that cause the performance anomaly or bottleneck problem, understand the causes and consequences of the performance anomaly more comprehensively, locate the problem more effectively, optimize the system configuration, and improve the overall performance and stability of the storage system.

[0159] In a specific implementation of the present invention, step S225 may include the following steps:

[0160] Step 1: When the absolute value of the difference between the correlation coefficient and 1 is less than a first preset value, it is determined that the first performance indicator is positively correlated with the second performance indicator;

[0161] Step 2: When the absolute value of the difference between the correlation coefficient and -1 is less than a second preset value, it is determined that the first performance indicator is negatively correlated with the second performance indicator;

[0162] Step three: When the absolute value of the difference between the correlation coefficient and 0 is less than a third preset value, it is determined that the first performance indicator is independent of the second performance indicator.

[0163] For the convenience of description, the above three steps can be combined for explanation.

[0164] After calculating the correlation coefficient between the first indicator sequence and the second indicator sequence composed of the screened second performance indicators, when the absolute value of the difference between the correlation coefficient and 1 is less than the first preset value, it is determined that the first performance indicator is positively correlated with the second performance indicator, when the absolute value of the difference between the correlation coefficient and -1 is less than the second preset value, it is determined that the first performance indicator is negatively correlated with the second performance indicator, and when the absolute value of the difference between the correlation coefficient and 0 is less than the third preset value, it is determined that the first performance indicator is irrelevant to the second performance indicator. Thus, the correlation coefficient can accurately characterize the correlation between the performance indicators, providing convenience for subsequent abnormality troubleshooting.

[0165] It should be noted that the first preset value, the second preset value and the third preset value can be set and adjusted according to actual conditions, and the embodiment of the present invention does not limit this. The first, second and third of the first preset value, the second preset value and the third preset value have no distinction in size or order, and are only for distinguishing the thresholds corresponding to positive correlation, negative correlation and indicator irrelevance.

[0166] In a specific implementation of the present invention, after obtaining each performance indicator subset, the method may further include the following steps:

[0167] Step 1: Obtain the maximum performance index in each performance index subset respectively;

[0168] Step 2: Use the mean value and the maximum performance index to record each performance index subset.

[0169] For the convenience of description, the above two steps can be combined for explanation.

[0170] After obtaining each performance indicator subset, the maximum performance indicator in each performance indicator subset is obtained respectively, and each performance indicator subset is recorded using the performance indicator mean and the maximum performance indicator respectively. By recording each performance indicator subset using the performance indicator mean and the maximum performance indicator, a concise representation of each performance indicator subset is achieved.

[0171] Corresponding to the above method embodiment, the present invention further provides a storage system performance indicator anomaly detection device. The storage system performance indicator anomaly detection device described below and the storage system performance indicator anomaly detection method described above can refer to each other.

[0172] See also Figure 4 , Figure 4 This is a structural block diagram of a storage system performance indicator abnormality detection device in an embodiment of the present invention. The device may include:

[0173] A first performance indicator data set acquisition module 41 is used to acquire a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system;

[0174] A difference calculation module 42, used to respectively calculate the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set;

[0175] A data set segmentation module 43, used to segment the first performance indicator time series data set according to each difference value to obtain each performance indicator subset;

[0176] The performance indicator threshold calculation module 44 is used to calculate the performance indicator thresholds corresponding to each performance indicator subset;

[0177] The performance indicator anomaly detection module 45 is used to perform performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold.

[0178] It can be seen from the above technical solution that by respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set, the first performance indicator time series data set is segmented according to each difference, and an adaptive segmentation processing strategy is adopted. The segmentation length can be dynamically adjusted according to the characteristics of the data itself. Compared with the fixed-length segmentation method, it is more flexible and can more accurately adapt to the local changes of the time series, thereby providing a more accurate and dynamic threshold that is more adapted to the data characteristics. By comparing each first performance indicator in each performance indicator subset obtained by segmentation with the calculated performance indicator threshold, outliers can be accurately identified, accurate analysis of storage system performance indicators is achieved, and the accuracy of anomaly detection is improved.

[0179] In a specific embodiment of the present invention, the device may further include:

[0180] A normalization processing module is used to normalize each first performance indicator in the first performance indicator time series data set after obtaining the first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system, and before respectively calculating the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set.

[0181] In a specific embodiment of the present invention, the data set segmentation module includes:

[0182] The difference sequence acquisition submodule is used to sort the differences to obtain a difference sequence;

[0183] A target difference screening submodule is used to screen a preset number of target differences from the end of the difference sequence with a larger normalized difference;

[0184] A target performance indicator pair acquisition submodule is used to respectively acquire target performance indicator pairs of adjacent time series corresponding to each target difference value;

[0185] The data set segmentation submodule is used to segment the first performance indicator time series data set by taking the middle position of the first performance indicator of the previous time series and the first performance indicator of the next time series contained in each target performance indicator pair as the segmentation point.

[0186] In a specific implementation of the present invention, the performance indicator threshold calculation module includes:

[0187] The mean and standard deviation calculation submodule is used to calculate the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset respectively;

[0188] The performance indicator lower limit calculation submodule is used to calculate the difference between the mean of each performance indicator and three times the corresponding performance indicator standard deviation to obtain the performance indicator lower limit corresponding to each performance indicator subset;

[0189] The performance indicator upper limit calculation submodule is used to calculate the sum of the mean value of each performance indicator and three times the corresponding performance indicator standard deviation to obtain the performance indicator upper limit corresponding to each performance indicator subset;

[0190] The performance indicator threshold determination submodule is used to determine the performance indicator threshold corresponding to each performance indicator subset according to the performance indicator lower limit and the performance indicator upper limit corresponding to each performance indicator subset.

[0191] In a specific embodiment of the present invention, the device may further include:

[0192] A maximum performance indicator acquisition module is used to obtain the maximum performance indicator in each performance indicator subset after obtaining each performance indicator subset;

[0193] The performance indicator subset recording module is used to record each performance indicator subset using the performance indicator mean and the maximum performance indicator respectively.

[0194] In a specific implementation of the present invention, the performance indicator anomaly detection module includes:

[0195] A comparison result obtaining submodule is used to compare the first performance indicator in each performance indicator subset with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold value to obtain a comparison result;

[0196] The abnormal performance indicator determination submodule is used to determine that the first performance indicator is an abnormal performance indicator when the comparison result is less than the lower limit of the performance indicator or greater than the upper limit of the performance indicator.

[0197] In a specific embodiment of the present invention, the device may further include:

[0198] A judgment module, used to judge whether there are multiple continuous abnormal performance indicators;

[0199] A target time range determination module, used to determine the target time ranges corresponding to the multiple consecutive abnormal performance indicators when it is determined that there are multiple consecutive abnormal performance indicators;

[0200] The abnormal interval position determination module is used to determine the starting position and the ending position of the abnormal interval composed of multiple continuous abnormal performance indicators according to the target time range.

[0201] In a specific embodiment of the present invention, the device may further include:

[0202] A first indicator sequence screening module is used to screen a first indicator sequence to be detected for correlation from a first performance indicator time series data set according to the abnormal interval after determining the start position and the end position of the abnormal interval composed of a plurality of continuous abnormal performance indicators according to the target time range;

[0203] A second performance indicator data set acquisition module, used to acquire a second performance indicator time series data set to be detected for correlation;

[0204] A second performance indicator screening module, used to screen the second performance indicators corresponding to the time range to which the first indicator series belongs from the second performance indicator time series data set;

[0205] A correlation coefficient calculation module, used to calculate the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators;

[0206] The correlation determination module is used to determine the correlation between the first performance indicator and the second performance indicator according to the correlation coefficient.

[0207] In a specific embodiment of the present invention, the first indicator sequence screening module includes:

[0208] A first duration calculation submodule, used to calculate a first duration corresponding to a target time range;

[0209] A second duration obtaining submodule, used to divide the first duration into two equal parts to obtain a second duration;

[0210] A first performance indicator screening submodule, used to screen the first performance indicators corresponding to the first second duration of the target time range from the first performance indicator time series data set, and screen the first performance indicators corresponding to the second duration of the target time range;

[0211] The first indicator sequence determination submodule is used to determine the first performance indicators obtained by screening as a first indicator sequence.

[0212] In a specific embodiment of the present invention, the correlation determination module includes:

[0213] A positive correlation determination submodule, used to determine that the first performance indicator is positively correlated with the second performance indicator when the absolute value of the difference between the correlation coefficient and 1 is less than a first preset value;

[0214] A negative correlation determination submodule, used to determine that the first performance indicator is negatively correlated with the second performance indicator when the absolute value of the difference between the correlation coefficient and -1 is less than a second preset value;

[0215] The indicator-independence determination submodule is used to determine that the first performance indicator is independent of the second performance indicator when the absolute value of the difference between the correlation coefficient and 0 is less than a third preset value.

[0216] In a specific implementation of the present invention, the second performance indicator data set acquisition module includes:

[0217] A to-be-detected object determination submodule, used to determine the to-be-detected object to which the first performance indicator time series data set belongs from the storage system;

[0218] A target indicator type screening submodule is used to screen the target indicator type to be tested for correlation from the indicator type set corresponding to the object to be tested;

[0219] The second performance indicator data set determination submodule is used to determine a performance indicator time series data set of a target indicator type corresponding to the object to be detected as a second performance indicator time series data set.

[0220] In a specific embodiment of the present invention, the device may further include:

[0221] A first performance indicator acquisition module, configured to, when it is determined that there are missing values ​​in the first performance indicator time series data set, acquire a first performance indicator corresponding to a previous time series of a target time series to which the missing value belongs;

[0222] A module for determining a performance indicator to be filled, used to determine a first performance indicator corresponding to a previous time series of a target time series to which the acquired missing value belongs as a performance indicator to be filled;

[0223] The performance indicator filling module is used to fill the performance indicator to be filled into the corresponding position of the target time series in the first performance indicator time series data set.

[0224] Corresponding to the above method embodiment, see Figure 5 , Figure 5This is a schematic diagram of a storage system performance indicator anomaly detection device provided by the present invention, and the device may include:

[0225] A memory 332, for storing computer programs;

[0226] The processor 322 is used to implement the steps of the storage system performance indicator abnormality detection method of the above method embodiment when executing a computer program.

[0227] For details, please refer to Figure 6 , Figure 6 A specific structural diagram of a storage system performance indicator anomaly detection device provided for this embodiment, the storage system performance indicator anomaly detection device may have relatively large differences due to different configurations or performances, and may include a processor (central processing units, CPU) 322 (for example, one or more processors) and a memory 332, and the memory 332 stores one or more computer programs 342 or data 344. Among them, the memory 332 can be a temporary storage or a permanent storage. The program stored in the memory 332 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the data processing device. Furthermore, the processor 322 can be configured to communicate with the memory 332 to execute a series of instruction operations in the memory 332 on the storage system performance indicator anomaly detection device 301.

[0228] The storage system performance indicator abnormality detection device 301 may also include one or more power supplies 326 , one or more wired or wireless network interfaces 350 , one or more input and output interfaces 358 , and / or one or more operating systems 341 .

[0229] The steps in the storage system performance indicator anomaly detection method described above can be implemented by the structure of the storage system performance indicator anomaly detection device.

[0230] Corresponding to the above method embodiment, the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps can be implemented:

[0231] Obtain a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system; calculate the difference between the first performance indicators of each adjacent time series in the first performance indicator time series data set; segment the first performance indicator time series data set according to each difference to obtain each performance indicator subset; calculate the performance indicator threshold corresponding to each performance indicator subset; perform performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold.

[0232] The computer-readable storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0233] For an introduction to the computer-readable storage medium provided by the present invention, please refer to the above method embodiment, and the present invention will not be elaborated here.

[0234] Corresponding to the above method embodiment, the present invention further provides a computer program product, including a computer program, which implements the steps of the above storage system performance indicator abnormality detection method when executed by a processor.

[0235] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device, computer-readable storage medium, and computer program product disclosed in the embodiment, since they correspond to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0236] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the technical solution and core ideas of the present invention. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, the present invention can also be improved and modified, and these improvements and modifications also fall within the scope of protection of the present invention.

Claims

1. A method for detecting abnormality of storage system performance indicators, characterized in that: include: Acquire a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system; Respectively calculating the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set; Segmenting the first performance indicator time series data set according to each difference value to obtain each performance indicator subset; Calculate the performance indicator thresholds corresponding to each performance indicator subset; Perform performance indicator anomaly detection on each performance indicator subset according to each performance indicator threshold; Segmenting the first performance indicator time series data set according to the difference values ​​includes: Sort the differences by size to obtain a difference sequence; Filtering a preset number of target differences from the end with the larger difference in the difference sequence; Obtain target performance indicator pairs of adjacent time series corresponding to each target difference respectively; The first performance indicator time series data set is segmented by taking the middle position of the first performance indicator of the previous time series and the first performance indicator of the next time series contained in each target performance indicator pair as the segmentation point.

2. The storage system performance indicator abnormality detection method according to claim 1, characterized in that: After obtaining a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system, and before respectively calculating the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set, the method further includes: Normalization is performed on each first performance indicator in the first performance indicator time series data set.

3. The storage system performance indicator abnormality detection method according to claim 1, characterized in that: Calculate the performance indicator thresholds corresponding to each performance indicator subset, including: Calculate the performance indicator mean and performance indicator standard deviation corresponding to each performance indicator subset respectively; The difference between the mean of each performance indicator and three times the standard deviation of the corresponding performance indicator is calculated to obtain the lower limit of the performance indicator corresponding to each performance indicator subset; The mean of each performance indicator and three times the standard deviation of the corresponding performance indicator are summed up to obtain the upper limit of the performance indicator corresponding to each performance indicator subset; The performance indicator thresholds corresponding to each performance indicator subset are determined according to the performance indicator lower limits and performance indicator upper limits corresponding to each performance indicator subset.

4. The storage system performance indicator abnormality detection method according to claim 3, characterized in that: After obtaining each performance indicator subset, it also includes: Obtain the maximum performance index in each performance index subset respectively; Each performance indicator subset is recorded using the performance indicator mean and the maximum performance indicator respectively.

5. The storage system performance indicator abnormality detection method according to claim 3, characterized in that: According to the threshold of each performance indicator, each performance indicator subset is respectively subjected to performance indicator anomaly detection, including: Compare the first performance indicator in each performance indicator subset with the performance indicator lower limit and the performance indicator upper limit in the corresponding performance indicator threshold respectively to obtain a comparison result; When the comparison result is less than a lower limit of the performance indicator or greater than an upper limit of the performance indicator, it is determined that the first performance indicator is an abnormal performance indicator.

6. The storage system performance indicator abnormality detection method according to claim 5, characterized in that: After determining that the first performance indicator is an abnormal performance indicator, the method further includes: Determine whether there are multiple consecutive abnormal performance indicators; If so, determine the target time range corresponding to multiple consecutive abnormal performance indicators; The starting position and the ending position of the abnormal interval composed of a plurality of continuous abnormal performance indicators are determined according to the target time range.

7. The storage system performance indicator abnormality detection method according to claim 6, characterized in that: After determining the start position and the end position of the abnormal interval composed of a plurality of continuous abnormal performance indicators according to the target time range, the method further includes: Selecting a first indicator sequence to be detected for correlation from the first performance indicator time series data set according to the abnormal interval; Obtain a second performance indicator time series data set to be tested for correlation; Filtering each second performance indicator corresponding to the time range to which the first indicator series belongs from the second performance indicator time series data set; Calculating the correlation coefficient between the first indicator sequence and the second indicator sequence formed by the screened second performance indicators; The correlation between the first performance indicator and the second performance indicator is determined according to the correlation coefficient.

8. The storage system performance indicator abnormality detection method according to claim 7, characterized in that: Selecting a first indicator sequence to be detected for correlation from the first performance indicator time series data set according to the abnormal interval includes: Calculating a first duration corresponding to the target time range; Divide the first duration into two equal parts to obtain a second duration; Filter the first performance indicators corresponding to the first second duration of the target time range from the first performance indicator time series data set, and filter the first performance indicators corresponding to the second duration of the target time range; The first performance indicators obtained by screening are determined as the first indicator sequence.

9. The storage system performance indicator abnormality detection method according to claim 7, characterized in that: Determining the correlation between the first performance indicator and the second performance indicator according to the correlation coefficient includes: When the absolute value of the difference between the correlation coefficient and 1 is less than a first preset value, it is determined that the first performance indicator is positively correlated with the second performance indicator; When the absolute value of the difference between the correlation coefficient and -1 is less than a second preset value, it is determined that the first performance indicator is negatively correlated with the second performance indicator; When the absolute value of the difference between the correlation coefficient and 0 is smaller than a third preset value, it is determined that the first performance indicator is independent of the second performance indicator.

10. The storage system performance indicator abnormality detection method according to claim 7, characterized in that: Obtain a second performance indicator time series data set to be tested for correlation, including: Determine from the storage system the object to be detected to which the first performance indicator time series data set belongs; Filtering the target indicator type to be detected for correlation from the indicator type set corresponding to the object to be detected; A performance indicator time series data set of a target indicator type corresponding to the object to be detected is determined as the second performance indicator time series data set.

11. The storage system performance indicator abnormality detection method according to any one of claims 1 to 10, characterized in that: After obtaining a first performance indicator time series data set corresponding to the performance indicator to be detected in the storage system, and before respectively calculating the difference between the first performance indicators of adjacent time series in the first performance indicator time series data set, the method further includes: When it is determined that there are missing values ​​in the first performance indicator time series data set, obtaining a first performance indicator corresponding to a previous time series of the target time series to which the missing value belongs; Determine the first performance indicator corresponding to the previous time series of the target time series to which the missing value belongs as the performance indicator to be filled; Fill the performance indicator to be filled into the corresponding position of the target time series in the first performance indicator time series data set.

12. A storage system performance indicator abnormality detection device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the storage system performance indicator anomaly detection method as claimed in any one of claims 1 to 11 when executing the computer program.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the storage system performance indicator anomaly detection method according to any one of claims 1 to 11 are implemented.

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the storage system performance indicator anomaly detection method as described in any one of claims 1 to 11 are implemented.

Images