Gray version anomaly detection method and device, electronic equipment and storage medium
Through sliding window technology and historical data fitting, the problem of insufficient data on grayscale version monitoring indicators is solved, the accuracy of abnormal detection is improved, and the abnormal detection of grayscale version can be accurately performed.
Patent Information
- Application Number
- CN202411212383.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-12-06
AI Technical Summary
In the prior art, the application data of the grayscale version is small, which leads to the lack of sufficient data support for monitoring indicators and it is impossible to accurately determine whether there is an abnormality in the grayscale version.
By using a sliding window of preset length, the application data after the grayscale version is collected and aggregated, and the target monitoring indicators are fitted with the historical application data of the historical version, the data volume is expanded and the detection accuracy is improved.
The data support for grayscale version target monitoring indicators is sufficient, which improves the accuracy of abnormal detection and can more accurately determine whether there are abnormalities in grayscale versions.
Smart Images

Figure CN119088701A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a grayscale version of anomaly detection method, device, electronic device and storage medium. Background Art
[0002] In current software engineering, when it is necessary to update software or services that have already been applied online, a new version is generally developed to replace the old version that has already been launched. For the new version, there may be potential risks, such as problems that have not appeared in the code. Therefore, in order to prevent the potential risks in the new version from affecting the use of most users, gray release is generally chosen at present, that is, a part of users or a small number of servers are selected to release the new version first. Furthermore, the gray release version can be called a gray version.
[0003] Based on this, it is necessary to detect in real time whether there are any anomalies in the grayscale version during the application process, so that when there are anomalies in the grayscale version, it can be rolled back quickly with a smaller impact range, reducing the risk of releasing a new version.
[0004] At present, the application data of the gray version is generally counted, and whether the gray version is abnormal is determined based on the statistical monitoring indicators. However, due to the small application scope of the gray version, the application data of the gray version is small, which results in insufficient data support for the statistical monitoring indicators, making it impossible to accurately determine whether the gray version is abnormal. Summary of the invention
[0005] The present application provides a grayscale version anomaly detection method, device, electronic device and storage medium to solve the technical problem in the prior art that when counting monitoring indicators of the grayscale version, the statistical monitoring indicators do not have sufficient data support due to the small amount of application data of the grayscale version, resulting in the inability to accurately determine whether there is an anomaly in the grayscale version.
[0006] In a first aspect, the present application provides a grayscale version of anomaly detection method, the method comprising:
[0007] For any sliding window of a preset length, the application data after the gray version is released is collected and aggregated using the sliding window of the preset length to obtain the target monitoring index of the gray version collected each time by each sliding window;
[0008] Using historical application data of historical versions to fit standard monitoring indicators corresponding to target monitoring indicators collected by the sliding window, the historical versions include all versions before the gray version is released;
[0009] According to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator, anomaly detection is performed on the grayscale version.
[0010] As an optional implementation, the application data after the gray version is released is collected and aggregated using a sliding window of a preset length to obtain the target monitoring index of the gray version collected by each sliding window each time, including:
[0011] Determine a sliding interval and a sliding step size of a sliding window of a preset length;
[0012] Controlling the sliding window to slide on the application data after the grayscale version is released according to the sliding interval and the sliding step size of the sliding window;
[0013] The sliding window is used to collect and aggregate the application data covered by each movement of the sliding window to obtain the target monitoring index of the grayscale version collected by the sliding window each time.
[0014] As an optional implementation, the use of historical application data of the historical version to fit the standard monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window includes:
[0015] Determine the indicator type of the target monitoring indicator;
[0016] Determining, according to the indicator type, target historical application data corresponding to the grayscale version from the historical application data of the historical version;
[0017] The target historical application data is collected and aggregated using the sliding window to obtain an initial monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window;
[0018] The initial monitoring indicator is fitted using a pre-trained fitting model to obtain a standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time.
[0019] As an optional implementation, determining the target historical application data corresponding to the gray version from the historical application data of the historical version according to the indicator type includes:
[0020] When the indicator type is a ratio type, all the historical application data of the historical version are determined as the target historical application data corresponding to the grayscale version;
[0021] When the indicator type is data type, the number of grayscale users corresponding to the grayscale version and the number of historical users corresponding to the historical version are obtained;
[0022] Determining a user ratio of the grayscale version to the historical version according to the grayscale user quantity and the historical user quantity;
[0023] The historical application data of the historical version is scaled according to the user ratio to obtain the target historical application data corresponding to the grayscale version.
[0024] As an optional implementation, the target historical application data is collected and aggregated using the sliding window to obtain the initial monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window, including:
[0025] Determining a sliding interval and a sliding step size of the sliding window;
[0026] Controlling the sliding window to slide on the target historical application data according to the sliding interval and sliding step size of the sliding window;
[0027] The sliding window is used to collect and aggregate the target historical application data covered by each movement of the sliding window to obtain the initial monitoring index corresponding to the target monitoring index collected by the sliding window each time.
[0028] As an optional implementation, performing anomaly detection on the grayscale version according to the target monitoring indicator collected each time by each sliding window and the standard monitoring indicator corresponding to the target monitoring indicator includes:
[0029] For the target monitoring indicator collected each time in each sliding window, determine the standard indicator range corresponding to the target monitoring indicator according to the standard monitoring indicator corresponding to the target monitoring indicator and a preset anomaly detection algorithm;
[0030] Determine whether the target monitoring indicator collected by each sliding window each time is within the corresponding standard indicator range;
[0031] A sliding window in which the target monitoring indicator is outside the corresponding standard indicator range is determined as an abnormal sliding window;
[0032] A sliding window in which the target monitoring indicator is within the corresponding standard indicator range is determined as a non-abnormal sliding window;
[0033] Anomaly detection is performed on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window.
[0034] As an optional implementation, performing anomaly detection on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window includes:
[0035] Determining whether the grayscale version is abnormal according to the abnormal sliding window;
[0036] When it is determined according to the abnormal sliding window that the gray version does not have an abnormality, stop using the abnormal sliding window to collect and aggregate the application data after the gray version is released;
[0037] When it is determined that the grayscale version is abnormal according to the abnormal sliding window, updating the abnormal sliding window to a non-abnormal sliding window;
[0038] Using the non-abnormal sliding window to collect and aggregate the application data after the gray version is released, to obtain new target monitoring indicators collected by the non-abnormal sliding window;
[0039] For each of the non-abnormal sliding windows, the historical application data of the historical version is used to fit the new standard monitoring indicator corresponding to the new target monitoring indicator collected each time by the non-abnormal sliding window;
[0040] According to the new target monitoring indicator collected by each of the non-abnormal sliding windows and the new standard monitoring indicator corresponding to the new target monitoring indicator, anomaly detection is performed on the grayscale version.
[0041] In a second aspect, the present application provides a grayscale version of an abnormality detection device, the device comprising:
[0042] An indicator determination module is used to collect and aggregate application data after the grayscale version is released using a sliding window of a preset length for any sliding window of a preset length, and obtain a target monitoring indicator of the grayscale version collected each time by each sliding window;
[0043] An indicator fitting module, used to fit the standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window using the historical application data of the historical version, wherein the historical version includes all versions before the gray version is released;
[0044] The anomaly detection module is used to perform anomaly detection on the grayscale version according to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator.
[0045] As an optional implementation, the indicator determination module includes:
[0046] A determination submodule, used to determine a sliding interval and a sliding step size of a sliding window of a preset length;
[0047] A control submodule, used for controlling the sliding window to slide on the application data after the grayscale version is released according to the sliding interval and sliding step size of the sliding window;
[0048] The collection submodule is used to use the sliding window to collect and aggregate the application data covered by each movement of the sliding window to obtain the grayscale version of the target monitoring index collected by the sliding window each time.
[0049] As an optional implementation, the index fitting module includes:
[0050] An indicator type determination submodule, used to determine the indicator type of the target monitoring indicator;
[0051] A history determination submodule, used to determine the target historical application data corresponding to the gray version from the historical application data of the historical version according to the indicator type;
[0052] A collection and aggregation submodule, used to collect and aggregate the target historical application data using the sliding window to obtain an initial monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time;
[0053] The fitting submodule is used to fit the initial monitoring indicator using a pre-trained fitting model to obtain a standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time.
[0054] As an optional implementation, the history determination submodule is specifically used to:
[0055] When the indicator type is a ratio type, all the historical application data of the historical version are determined as the target historical application data corresponding to the grayscale version;
[0056] When the indicator type is data type, the number of grayscale users corresponding to the grayscale version and the number of historical users corresponding to the historical version are obtained;
[0057] Determining a user ratio of the grayscale version to the historical version according to the grayscale user quantity and the historical user quantity;
[0058] The historical application data of the historical version is scaled according to the user ratio to obtain the target historical application data corresponding to the grayscale version.
[0059] As an optional implementation, the collection and aggregation submodule is specifically used to:
[0060] Determining a sliding interval and a sliding step size of the sliding window;
[0061] Controlling the sliding window to slide on the target historical application data according to the sliding interval and sliding step size of the sliding window;
[0062] The sliding window is used to collect and aggregate the target historical application data covered by each movement of the sliding window to obtain the initial monitoring index corresponding to the target monitoring index collected by the sliding window each time.
[0063] As an optional implementation, the anomaly detection module includes:
[0064] The indicator range determination submodule is used to determine the standard indicator range corresponding to the target monitoring indicator collected each time in each sliding window according to the standard monitoring indicator corresponding to the target monitoring indicator and a preset anomaly detection algorithm;
[0065] The first determination submodule is used to determine whether the target monitoring indicator collected by each sliding window each time is within the corresponding standard indicator range;
[0066] The second determination submodule is used to determine a sliding window whose target monitoring indicator is outside the corresponding standard indicator range as an abnormal sliding window;
[0067] A third determination submodule is used to determine a sliding window whose target monitoring indicator is within the corresponding standard indicator range as a non-abnormal sliding window;
[0068] An anomaly detection submodule is used to perform anomaly detection on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window.
[0069] As an optional implementation, the anomaly detection submodule is specifically used to:
[0070] Determining whether the grayscale version is abnormal according to the abnormal sliding window;
[0071] When it is determined according to the abnormal sliding window that the gray version does not have an abnormality, stop using the abnormal sliding window to collect and aggregate the application data after the gray version is released;
[0072] When it is determined that the grayscale version is abnormal according to the abnormal sliding window, updating the abnormal sliding window to a non-abnormal sliding window;
[0073] Using the non-abnormal sliding window to collect and aggregate the application data after the gray version is released, to obtain new target monitoring indicators collected by the non-abnormal sliding window;
[0074] For each of the non-abnormal sliding windows, the historical application data of the historical version is used to fit the new standard monitoring indicator corresponding to the new target monitoring indicator collected each time by the non-abnormal sliding window;
[0075] According to the new target monitoring indicator collected by each of the non-abnormal sliding windows and the new standard monitoring indicator corresponding to the new target monitoring indicator, anomaly detection is performed on the grayscale version.
[0076] In a third aspect, the present application provides an electronic device comprising: a processor and a memory, wherein the processor is used to execute a grayscale version of anomaly detection program stored in the memory to implement the grayscale version anomaly detection method described in any one of the first aspects.
[0077] In a fourth aspect, the present application provides a storage medium storing one or more programs, which can be executed by one or more processors to implement the grayscale version of the anomaly detection method described in any one of the first aspects.
[0078] The technical solution provided by the embodiment of the present application is to collect and aggregate the application data after the gray version is released by using a sliding window of a preset length for any sliding window, obtain the target monitoring index of the gray version collected by each sliding window each time, and use the historical application data of the historical version to fit the standard monitoring index corresponding to the target monitoring index collected by the sliding window. The above-mentioned historical version includes all versions before the gray version is released. According to the target monitoring index collected by each sliding window each time and the standard monitoring index corresponding to the target monitoring index, the gray version is detected for abnormality. This technical solution rolls and aggregates the application data of the gray version through sliding windows of different lengths, and rolls and aggregates the application data through sliding windows of different lengths, and the amount of application data can be expanded by using the sliding window, and each sliding window can collect the target monitoring index of the gray version each time, and finally multiple target monitoring indicators can be obtained. At the same time, the historical application data of the historical version is used as the historical application data of the gray version, and the standard monitoring index of each target monitoring index is fitted, so as to more accurately determine whether the gray version is abnormal based on more target monitoring indicators and standard monitoring indicators, so as to achieve sufficient data support for the target monitoring index of the collected gray version, thereby achieving more accurate determination of whether the gray version is abnormal. BRIEF DESCRIPTION OF THE DRAWINGS
[0079] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0080] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art texts are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0081] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.
[0082] Figure 1 A flowchart of an embodiment of a grayscale version of an abnormality detection method provided in an embodiment of the present application;
[0083] Figure 2 A flowchart of another grayscale version of an abnormality detection method provided in an embodiment of the present application;
[0084] Figure 3 A flowchart of another grayscale version of an abnormality detection method provided in an embodiment of the present application;
[0085] Figure 4 A block diagram of an embodiment of a grayscale version of an abnormality detection device provided in an embodiment of the present application;
[0086] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0087] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the embodiments described in the text are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0088] The disclosure below provides many different embodiments or examples to implement different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.
[0089] In order to solve the technical problem in the prior art that when counting monitoring indicators of grayscale versions, the statistical monitoring indicators do not have sufficient data support due to the small amount of application data of the grayscale version, thereby making it impossible to accurately determine whether the grayscale version has an abnormality, the present application provides a grayscale version anomaly detection method, device, electronic device and storage medium, which can perform rolling aggregation on the application data of the grayscale version through sliding windows of different lengths. By performing rolling aggregation on the application data through sliding windows of different lengths, the sliding window can be used to expand the data volume of the application data, and each sliding window can collect the target monitoring indicators of the grayscale version each time, and finally multiple target monitoring indicators can be obtained. At the same time, the historical application data of the historical version is used as the historical application data of the grayscale version, and the standard monitoring indicator of each target monitoring indicator is fitted, so as to more accurately determine whether the grayscale version has an abnormality based on more target monitoring indicators and standard monitoring indicators, so that the collected target monitoring indicators of the grayscale version have sufficient data support, thereby more accurately determining whether the grayscale version has an abnormality.
[0090] The grayscale version of the anomaly detection method provided in the present application is further explained below with reference to specific embodiments in conjunction with the accompanying drawings. The embodiments do not constitute a limitation on the embodiments of the present invention.
[0091] See also Figure 1 , is a flowchart of an embodiment of a grayscale version of anomaly detection method provided in an embodiment of the present application. Figure 1 As shown, the process may include the following steps:
[0092] Step 101: For any sliding window of preset length, use the sliding window of preset length to collect and aggregate application data after the gray version is released, and obtain the target monitoring index of the gray version collected by each sliding window each time.
[0093] The above sliding window is a sliding time window, and its corresponding preset length is a preset time length. In this application, multiple sliding windows with different preset lengths can be preset, and the above sliding window is any sliding window among the multiple sliding windows.
[0094] The above-mentioned grayscale version refers to the application or service version released in the grayscale stage.
[0095] The above-mentioned target monitoring indicators refer to indicators that can be used to measure whether the gray version is abnormal, such as the crash rate, failure rate, number of crashes, or number of failures used by users during the operation of the gray version.
[0096] The above-mentioned application data refers to the application data collected when the grayscale version is running and the user applies the grayscale version. It may be log data collected when the user applies the grayscale version, or it may be statistical data collected when the grayscale version is applied by the user after the grayscale version is released. The embodiments of the present application do not impose any restrictions on this.
[0097] In an embodiment of the present application, in order to count the monitoring indicators of the grayscale version during the application process and to ensure that the monitoring indicators have sufficient data support, the executing entity of the embodiment of the present application can use pre-set sliding windows of different preset lengths to count the application data after the grayscale version is released, thereby obtaining the monitoring indicators of the grayscale version during the application process.
[0098] In one embodiment, the execution subject of the embodiment of the present application can collect and aggregate the application data after the grayscale version is released by using any sliding window of a preset length among a plurality of sliding windows of different preset lengths, and obtain the target monitoring index of the grayscale version collected by each sliding window each time. According to the above method, each sliding window can collect and aggregate the corresponding target monitoring index after each movement.
[0099] Among them, since the grayscale version targets a smaller user range, the amount of application data obtained is smaller. If the preset length of the sliding window is shorter, the collected application data may have breakpoints. Therefore, in the above-mentioned pre-set sliding windows of different preset lengths, the preset length of each sliding window may be greater than or equal to the preset time length threshold (for example, 1 minute). The preset length of the above-mentioned sliding window may be 1 minute, 10 minutes, 3 minutes, or 1 hour, and the embodiments of the present application do not impose any restrictions on this.
[0100] As an optional implementation, the execution subject of the embodiment of the present application can obtain the sliding interval and sliding step of the sliding window of the preset length when determining the target monitoring index of the grayscale version collected by each sliding window each time. The above sliding interval refers to the time interval between two movements of the sliding window, and the above sliding step refers to the length of each sliding of the sliding window. The length can be the length of the storage space occupied by the application data or the data length, and the embodiment of the present application does not limit this.
[0101] Afterwards, the sliding window can be controlled to slide on the application data after the grayscale version is released according to the sliding interval and sliding step of the above-mentioned sliding window, and the sliding window can be used to collect and aggregate the application data covered by each movement of the sliding window to obtain the target monitoring indicators of the grayscale version collected by the sliding window each time.
[0102] As an exemplary implementation, the sliding window may first collect the application data covered, and aggregate the application data according to the target monitoring index, thereby obtaining the target monitoring index of the grayscale version. Aggregation here refers to analysis, classification, or calculation of the application data.
[0103] For example, assuming that the target monitoring indicator is the failure rate, the sliding window can collect the covered application data, and count the failed users of the grayscale version of the application in the application data, as well as the total users, and then calculate the failure rate based on the above failed users and total users, so as to obtain the target monitoring indicator of the sliding window collected this time.
[0104] Step 102: Use the historical application data of the historical version to fit the standard monitoring indicators corresponding to the target monitoring indicators collected by the sliding window. The above historical versions include all versions before the gray version is released.
[0105] The above-mentioned historical versions refer to the versions released before the grayscale version, which are all versions of the same application or service.
[0106] The above-mentioned standard monitoring indicators refer to the monitoring indicators when there is no abnormality in the grayscale version fitted according to the historical application data of the grayscale version. Among them, since the grayscale version is the latest developed version, there is no historical application data. Therefore, in the embodiment of the present application, the application data of the historical version corresponding to the grayscale version can be used as the historical application data of the grayscale version.
[0107] In an embodiment of the present application, in order to determine whether there is an abnormality in the grayscale version based on the target monitoring indicators of the collected grayscale version, the executing entity of the embodiment of the present application can use the historical application data of the historical version to fit the standard monitoring indicators corresponding to the target monitoring indicators collected each time by the sliding window, thereby using it as a standard for measuring whether there is an abnormality in the grayscale version.
[0108] As for how to use the historical application data of the historical version to fit the standard monitoring indicators corresponding to the target monitoring indicators collected by the sliding window, it can be seen in the following Figure 2 The process shown is explained and will not be described in detail here.
[0109] Step 103: Perform anomaly detection on the grayscale version according to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator.
[0110] In an embodiment of the present application, after determining the target monitoring indicators collected by each sliding window each time and the standard monitoring indicators corresponding to each target monitoring indicator according to the above steps, anomaly detection can be performed on the grayscale version based on each target monitoring indicator and the standard monitoring indicator corresponding to the target monitoring indicator.
[0111] As an optional implementation, for each target monitoring indicator collected by each sliding window each time, the standard indicator range of the target monitoring indicator can be determined according to the standard monitoring indicator corresponding to the target monitoring indicator and a preset anomaly detection algorithm. The above anomaly detection algorithm can be a commonly used anomaly detection algorithm such as 3-Sigma.
[0112] As an exemplary implementation, when determining the standard indicator range of the target monitoring indicator according to the standard monitoring indicator corresponding to the target monitoring indicator and the preset anomaly detection algorithm, the average value and standard deviation of the standard monitoring indicator can be determined. Then, the standard deviation can be multiplied by the preset value to obtain the standard value, and then the standard value can be subtracted from the average value to obtain the first value, and the average value can be added to the standard value to obtain the second value, and the range consisting of the first value and the second value is determined as the standard indicator range.
[0113] Afterwards, it can be determined whether the target monitoring indicator is within the corresponding standard indicator range.
[0114] Optionally, when it is determined that the target monitoring indicator is within the corresponding standard indicator range, it can be determined that there is no abnormality in the grayscale version represented by the target monitoring indicator.
[0115] On the contrary, when it is determined that the target monitoring indicator is outside the corresponding standard indicator range, it can be determined that the target monitoring indicator represents that the grayscale version is abnormal.
[0116] Furthermore, in order to exclude the possibility that the target monitoring indicator is not within the standard indicator range due to the small amount of application data, the executing entity of the embodiment of the present application can output the target monitoring indicator after determining that the target monitoring indicator is not within the standard indicator range, so that the technical staff can determine whether there is an abnormality in the grayscale version.
[0117] As another optional implementation method, in order to improve the accuracy of anomaly detection for grayscale versions, the executing entity of the embodiment of the present application may exclude sliding windows with inaccurate target monitoring indicators, and subsequently use sliding windows with accurate target monitoring indicators to perform anomaly detection on grayscale versions. The sliding window with relatively accurate collected target monitoring indicators indicates that the target monitoring indicators collected during the sliding process have better continuity. Therefore, using this type of sliding window to perform anomaly detection on grayscale versions in the subsequent process is also more accurate.
[0118] As an exemplary implementation, for each target monitoring indicator collected each time in each sliding window, the standard indicator range of the target monitoring indicator can be determined according to the standard monitoring indicator corresponding to the target monitoring indicator and the preset anomaly detection algorithm. The above anomaly detection algorithm can be a commonly used anomaly detection algorithm such as 3-Sigma. Afterwards, it can be determined whether the target monitoring indicator is within the corresponding standard indicator range.
[0119] Optionally, a sliding window where the target monitoring indicator is outside the corresponding standard range can be determined as an abnormal sliding window, and a sliding window where the target monitoring indicator is within the corresponding standard indicator range can be determined as a non-abnormal sliding window. Based on this, anomaly detection can be performed on the grayscale version according to the above abnormal sliding windows and non-abnormal sliding windows.
[0120] As an exemplary implementation, when anomaly detection is performed on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window, it is possible to determine whether the grayscale version is abnormal according to the abnormal sliding window. Among them, the technician can manually determine whether the grayscale version abnormality indicated by the abnormal sliding window is accurate, or the pre-trained abnormality detection model can be used to determine whether the grayscale version abnormality indicated by the abnormal sliding window is accurate, and the embodiments of the present application do not limit this.
[0121] Optionally, when it is determined that there is no abnormality in the grayscale version based on the abnormal sliding window, it means that the target monitoring indicators collected and aggregated by the abnormal sliding window are inaccurate. Therefore, the collection and aggregation of application data after the grayscale version is released using the abnormal sliding window can be stopped.
[0122] Optionally, when it is determined that an abnormality exists in the grayscale version based on the abnormal sliding window, it means that the target monitoring indicators collected and aggregated by the abnormal sliding window are relatively accurate, so the abnormal sliding window can be updated to a non-abnormal sliding window.
[0123] Furthermore, the execution subject of the embodiment of the present application can use the current non-abnormal sliding window to collect and aggregate the application data after the grayscale version is released, and obtain new target monitoring indicators collected by the non-abnormal sliding window.
[0124] Afterwards, for each non-abnormal sliding window, the historical application data of the historical version can be used to fit the new standard monitoring indicator corresponding to the new target monitoring indicator collected each time in the non-abnormal sliding window. The method of fitting the new standard monitoring indicator is similar to the method of fitting the standard monitoring indicator in step 102, and will not be described in detail here.
[0125] Finally, the gray version can be detected for anomalies based on the new target monitoring indicators collected by each non-abnormal sliding window and the new standard monitoring indicators corresponding to the new target monitoring indicators. The method for detecting anomalies of the gray version based on the new target monitoring indicators and the new standard monitoring indicators is similar to the method for detecting anomalies of the gray version based on the target monitoring indicators and the standard monitoring indicators in step 103, and will not be described in detail here.
[0126] The technical solution provided by the embodiment of the present application is to collect and aggregate the application data after the gray version is released by using a sliding window of a preset length for any sliding window, obtain the target monitoring index of the gray version collected by each sliding window each time, and use the historical application data of the historical version to fit the standard monitoring index corresponding to the target monitoring index collected by the sliding window. The above-mentioned historical version includes all versions before the gray version is released. According to the target monitoring index collected by each sliding window each time and the standard monitoring index corresponding to the target monitoring index, the gray version is detected for abnormality. This technical solution rolls and aggregates the application data of the gray version through sliding windows of different lengths, and rolls and aggregates the application data through sliding windows of different lengths, and the amount of application data can be expanded by using the sliding window, and each sliding window can collect the target monitoring index of the gray version each time, and finally multiple target monitoring indicators can be obtained. At the same time, the historical application data of the historical version is used as the historical application data of the gray version, and the standard monitoring index of each target monitoring index is fitted, so as to more accurately determine whether the gray version is abnormal based on more target monitoring indicators and standard monitoring indicators, so as to achieve sufficient data support for the target monitoring index of the collected gray version, thereby achieving more accurate determination of whether the gray version is abnormal.
[0127] See also Figure 2 , which is a flow chart of an embodiment of another grayscale version of anomaly detection method provided in an embodiment of the present application. Figure 2 The process shown in Figure 1 Based on the process shown in FIG, it describes how to use the historical application data of the historical version to fit the standard monitoring indicators corresponding to the target monitoring indicators collected by the sliding window. Figure 2 As shown, the process may include the following steps:
[0128] Step 201: Determine the indicator type of the target monitoring indicator.
[0129] Step 202: According to the above indicator type, determine the target historical application data corresponding to the gray version from the historical application data of the historical version.
[0130] The following is a unified description of step 201 and step 202:
[0131] The above indicator types refer to the statistical types represented by the target monitoring indicators, such as proportion type and data type. The proportion type means that the statistics are calculated according to the proportion, such as failure rate and crash rate, etc. The data type means that the statistical quantity is directly calculated, such as the number of failures and crashes.
[0132] In the embodiment of the present application, since there is no historical application data after the grayscale version is released, the target historical application data corresponding to the grayscale version can be determined from the historical application data of the historical version.
[0133] Furthermore, the execution subject of the embodiment of the present application may determine the indicator type of the target monitoring indicator, and determine the target historical application data corresponding to the grayscale version from the historical application data of the historical version according to the indicator type.
[0134] As an optional implementation method, the indicator type of the target monitoring indicator may include a proportional type and a data type. Based on this, when the executing entity of the embodiment of the present application determines that the indicator type of the target monitoring indicator is a proportional type, it can determine all the historical application data of the historical version as the target historical application data corresponding to the grayscale version.
[0135] Optionally, when it is determined that the indicator type of the target monitoring indicator is a data type, the number of users corresponding to the grayscale version (for ease of description, hereinafter referred to as the grayscale user quantity) and the number of users corresponding to the historical version (for ease of description, hereinafter referred to as the historical user quantity) can be obtained.
[0136] Afterwards, the user ratio of the gray version and the historical version can be determined based on the gray user volume and historical user volume, and the historical application data of the historical version can be scaled based on the user ratio to obtain the target historical application data corresponding to the gray version.
[0137] Among them, the ratio of the above-mentioned grayscale user volume to the historical user volume can be determined as the user ratio of the grayscale version and the historical version.
[0138] As an exemplary implementation, when scaling the historical application data of the historical version according to the user ratio, the target historical application data may be randomly selected from the historical application data of the historical version according to the user ratio.
[0139] As another exemplary implementation, when scaling the historical application data of the historical version according to the user ratio, the historical application data of the historical version can be sorted according to the order of the generation time of the historical application data to obtain a historical application data sequence. Afterwards, the target historical application data with the most recent generation time can be selected from the historical application data sequence according to the above user ratio in the order of the generation time from the latest to the earliest.
[0140] Step 203: Use the sliding window to collect and aggregate the target historical application data to obtain the initial monitoring index corresponding to the target monitoring index collected by the sliding window each time.
[0141] Step 204: fit the initial monitoring index using the pre-trained fitting model to obtain the standard monitoring index corresponding to the target monitoring index collected each time by the sliding window.
[0142] The following is a unified description of step 203 and step 204:
[0143] The above-mentioned initial monitoring indicator refers to the initial monitoring indicator corresponding to the target monitoring indicator obtained after the sliding window collects and aggregates the target historical application data. The initial monitoring indicator and the target monitoring indicator are monitoring indicators of different application data in the same time period collected by the same sliding window. The initial monitoring indicator targets the target historical application data selected from the historical application data of the historical version, and the target monitoring indicator targets the application data after the gray version is released.
[0144] The above-mentioned standard monitoring indicators refer to the standard monitoring indicators corresponding to each target monitoring indicator when there is no abnormality in the grayscale version fitted by the pre-trained fitting model.
[0145] In one embodiment, after selecting the target historical application data corresponding to the grayscale version from the historical application data of the historical version, when determining the standard monitoring indicator corresponding to each target monitoring indicator, the target historical application data can be first collected and aggregated using the sliding window corresponding to the target monitoring indicator to obtain the initial monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window.
[0146] As an optional implementation method, the sliding interval and sliding step of the sliding window can be determined. The above sliding interval refers to the time interval between two movements of the sliding window, and the above sliding step refers to the length of each sliding of the sliding window. The length can be the length of the storage space occupied by the application data or the data length. The embodiment of the present application does not impose any restrictions on this.
[0147] Afterwards, the sliding window can be controlled to slide on the above-mentioned target historical application data according to the sliding interval and sliding step size of the above-mentioned sliding window, and the sliding window can be used to collect and aggregate the target historical application data covered by each movement of the sliding window to obtain the initial monitoring indicators corresponding to the target monitoring indicators collected by the sliding window each time.
[0148] Furthermore, since the target historical application data is selected from the historical application data of the historical version, and the initial monitoring indicator is obtained by collecting and aggregating the target historical application data with a sliding window, the initial monitoring indicator can be the monitoring indicator of the historical version during the application process.
[0149] Based on this, in order to further determine the standard monitoring indicators of the grayscale version, the execution subject of the embodiment of the present application can use the pre-trained fitting model to fit the initial monitoring indicators to obtain the standard monitoring indicators corresponding to the target monitoring indicators collected by the sliding window each time. The above-mentioned fitting model can be a pre-trained model for fitting the standard monitoring indicators corresponding to the target monitoring indicators, and can be an LSTM (Long Short-Term Memory, long short-term memory network) model, or other models, which is not limited by the embodiment of the present application.
[0150] The technical solution provided by the embodiment of the present application determines the indicator type of the target monitoring indicator, determines the target historical application data corresponding to the grayscale version from the historical application data of the historical version according to the above indicator type, collects and aggregates the target historical application data using a sliding window, obtains the initial monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window, and fits the initial monitoring indicator using a pre-trained fitting model to obtain the standard monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window. This technical solution uses the historical application data of the historical version as the historical application data of the grayscale version, and fits the standard indicator corresponding to the target monitoring indicator collected each time by the sliding window on this basis, thereby judging whether the grayscale version has an abnormality, and realizes fitting the standard indicator of the grayscale version, thereby more accurately determining whether the grayscale version has an abnormality.
[0151] See also Figure 3 , is a flowchart of another grayscale version of anomaly detection method provided in the present application. Figure 3 As shown, the process may include the following steps:
[0152] In the embodiment of the present application, in order to ensure the real-time monitoring of the target monitoring indicators (such as crash rate and failure rate) of the gray version, during monitoring, observation and analysis are generally performed at a relatively small time granularity. However, when the gray version is released, the number of users of the gray version is very small. Simply observing with a smaller time granularity (such as 1 minute) may result in breakpoints, and the regularity of the core indicators cannot be observed. Therefore, in the embodiment of the present application, when performing anomaly detection on the gray version, the execution subject of the embodiment of the present application can collect and aggregate the application data of the gray version through a variety of sliding windows, wherein the length of the window is relatively large (such as 1 minute, 10 minutes, 30 minutes, 1 hour...), and its real-time performance remains consistent (such as 1 minute) for rolling aggregation, so that the amount of data of the observed target monitoring indicators is sufficient to observe the regularity.
[0153] At the same time, the release of a grayscale version means that there is no historical data, and it is impossible to use historical monitoring indicator data to detect whether the current target monitoring indicator has any abnormalities. To address this issue, in an embodiment of the present application, the overall indicator (that is, the historical application data of the historical version) can be directly used to perform real-time baseline estimation based on the real-time grayscale situation (that is, estimating the standard monitoring indicator for the target monitoring indicator). On this basis, through traditional anomaly detection methods, such as 3-sigma and other statistical methods, it is possible to quickly determine whether the grayscale has abnormalities, and then decide whether to continue the grayscale.
[0154] Furthermore, the specific implementation steps are as follows:
[0155] 1. Multi-time window sliding aggregation monitoring settings:
[0156] Initialize sliding windows of different sizes, such as 1 minute, 10 minutes, 30 minutes, 1 hour, etc. Slide once every certain period of time (such as 1 minute), collect application data in the corresponding time period, and generate target monitoring indicators.
[0157] 2. Real-time baseline (standard monitoring indicators) estimation:
[0158] Use the overall indicator data (historical application data of historical versions) to build a real-time baseline model (also known as a fitting model). By analyzing the historical application data of the historical version before the gray version is released, if it is a proportional target monitoring indicator (such as crash rate, error rate), directly use the data obtained by sliding aggregation of historical application data as the target historical application data; if it is a numerical target monitoring indicator (such as the number of crashes, the number of errors), then according to the user volume, scale the overall same target monitoring indicator proportionally and perform sliding aggregation as the target historical application data, and then use LSTM or other fitting models to fit the standard monitoring indicator corresponding to the target monitoring indicator collected each time in each sliding window.
[0159] 3. Anomaly Detection Algorithm:
[0160] Commonly used anomaly detection algorithms such as 3-Sigma can be used (set a normal operating range (also known as the standard indicator range) for each window, such as mean (μ) ± 3*standard deviation (σ). If the target indicator of the grayscale version exceeds this range, it is considered abnormal).
[0161] Furthermore, for the sliding window with abnormal target monitoring indicators, if it is determined that there is no abnormality in the grayscale version, it means that the target monitoring indicators of the sliding window are inaccurate, and therefore the sliding window can be excluded; for the sliding window with no abnormality in the target monitoring indicators, it means that the continuity of the target monitoring indicators collected by the sliding window is good, and therefore, the sliding window with good continuity can be used to perform anomaly detection on the grayscale version in the future.
[0162] The technical solution provided in the embodiment of the present application, in order to address the problem of small user volume of the grayscale version, aggregates core monitoring indicators from the perspective of multiple sliding windows of different lengths, so that the laws of the core indicators can be observed. At the same time, in order to address the problem of lack of historical application data for the grayscale version, historical application data is fitted through overall data, so that the target monitoring indicators have historical application data as a reference, thereby accurately detecting anomalies in the grayscale version.
[0163] See also Figure 4 , is a block diagram of an embodiment of a grayscale version of an abnormality detection device provided in an embodiment of the present application. Figure 4 As shown, the device may include:
[0164] The indicator determination module 41 is used for collecting and aggregating application data after the grayscale version is released using a sliding window of a preset length for any sliding window of a preset length, and obtaining a target monitoring indicator of the grayscale version collected each time by each sliding window;
[0165] An indicator fitting module 42, used to fit the standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window using the historical application data of the historical version, wherein the historical version includes all versions before the gray version is released;
[0166] The anomaly detection module 43 is used to perform anomaly detection on the grayscale version according to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator.
[0167] As an optional implementation, the indicator determination module 41 includes:
[0168] A determination submodule, used to determine a sliding interval and a sliding step size of a sliding window of a preset length;
[0169] A control submodule, used for controlling the sliding window to slide on the application data after the grayscale version is released according to the sliding interval and sliding step size of the sliding window;
[0170] The collection submodule is used to use the sliding window to collect and aggregate the application data covered by each movement of the sliding window to obtain the grayscale version of the target monitoring index collected by the sliding window each time.
[0171] As an optional implementation, the index fitting module 42 includes:
[0172] An indicator type determination submodule, used to determine the indicator type of the target monitoring indicator;
[0173] A history determination submodule, used to determine the target historical application data corresponding to the gray version from the historical application data of the historical version according to the indicator type;
[0174] A collection and aggregation submodule, used to collect and aggregate the target historical application data using the sliding window to obtain an initial monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time;
[0175] The fitting submodule is used to fit the initial monitoring indicator using a pre-trained fitting model to obtain a standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time.
[0176] As an optional implementation, the history determination submodule is specifically used to:
[0177] When the indicator type is a ratio type, all the historical application data of the historical version are determined as the target historical application data corresponding to the grayscale version;
[0178] When the indicator type is data type, the number of grayscale users corresponding to the grayscale version and the number of historical users corresponding to the historical version are obtained;
[0179] Determining a user ratio of the grayscale version to the historical version according to the grayscale user quantity and the historical user quantity;
[0180] The historical application data of the historical version is scaled according to the user ratio to obtain the target historical application data corresponding to the grayscale version.
[0181] As an optional implementation, the collection and aggregation submodule is specifically used to:
[0182] Determining a sliding interval and a sliding step size of the sliding window;
[0183] Controlling the sliding window to slide on the target historical application data according to the sliding interval and sliding step size of the sliding window;
[0184] The sliding window is used to collect and aggregate the target historical application data covered by each movement of the sliding window to obtain the initial monitoring index corresponding to the target monitoring index collected by the sliding window each time.
[0185] As an optional implementation, the anomaly detection module 43 includes:
[0186] The indicator range determination submodule is used to determine the standard indicator range corresponding to the target monitoring indicator collected each time in each sliding window according to the standard monitoring indicator corresponding to the target monitoring indicator and a preset anomaly detection algorithm;
[0187] The first determination submodule is used to determine whether the target monitoring indicator collected by each sliding window each time is within the corresponding standard indicator range;
[0188] The second determination submodule is used to determine a sliding window whose target monitoring indicator is outside the corresponding standard indicator range as an abnormal sliding window;
[0189] A third determination submodule is used to determine a sliding window whose target monitoring indicator is within the corresponding standard indicator range as a non-abnormal sliding window;
[0190] An anomaly detection submodule is used to perform anomaly detection on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window.
[0191] As an optional implementation, the anomaly detection submodule is specifically used to:
[0192] Determining whether the grayscale version is abnormal according to the abnormal sliding window;
[0193] When it is determined according to the abnormal sliding window that the gray version does not have an abnormality, stop using the abnormal sliding window to collect and aggregate the application data after the gray version is released;
[0194] When it is determined that the grayscale version is abnormal according to the abnormal sliding window, updating the abnormal sliding window to a non-abnormal sliding window;
[0195] Using the non-abnormal sliding window to collect and aggregate the application data after the gray version is released, to obtain new target monitoring indicators collected by the non-abnormal sliding window;
[0196] For each of the non-abnormal sliding windows, the historical application data of the historical version is used to fit the new standard monitoring indicator corresponding to the new target monitoring indicator collected each time by the non-abnormal sliding window;
[0197] According to the new target monitoring indicator collected by each of the non-abnormal sliding windows and the new standard monitoring indicator corresponding to the new target monitoring indicator, anomaly detection is performed on the grayscale version.
[0198] like Figure 5 As shown, it is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, including a processor 51, a communication interface 52, a memory 53 and a communication bus 54, wherein the processor 51, the communication interface 52, and the memory 53 communicate with each other through the communication bus 54.
[0199] A memory 53, for storing computer programs;
[0200] In one embodiment of the present application, the processor 51 is used to execute the program stored in the memory 53 to implement the grayscale version of the abnormality detection method provided by any of the above method embodiments, including:
[0201] For any sliding window of a preset length, the application data after the gray version is released is collected and aggregated using the sliding window of the preset length to obtain the target monitoring index of the gray version collected each time by each sliding window;
[0202] Using historical application data of historical versions to fit standard monitoring indicators corresponding to target monitoring indicators collected by the sliding window, the historical versions include all versions before the gray version is released;
[0203] According to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator, anomaly detection is performed on the grayscale version.
[0204] An embodiment of the present application further provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the grayscale version of the anomaly detection method provided in any of the aforementioned method embodiments are implemented.
[0205] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.
[0206] Through the text of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.
[0207] It should be understood that the terms used in the text are only for the purpose of the specific example implementation of the text, and are not intended to be limited. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used in the text can also be represented to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of the stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or their combinations. The method steps, processes, and operations of the text in the text are not interpreted as necessarily requiring them to be performed in the specific order of the text or description, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps can be used.
[0208] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.
Claims
1. A grayscale version of anomaly detection method, characterized in that: The method comprises: For any sliding window of a preset length, the application data after the gray version is released is collected and aggregated using the sliding window of the preset length to obtain the target monitoring index of the gray version collected each time by each sliding window; Using historical application data of historical versions to fit standard monitoring indicators corresponding to target monitoring indicators collected by the sliding window, the historical versions include all versions before the gray version is released; According to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator, anomaly detection is performed on the grayscale version.
2. The method according to claim 1, characterized in that The use of a sliding window of a preset length to collect and aggregate application data after the grayscale version is released, and obtaining target monitoring indicators of the grayscale version collected by each sliding window each time, includes: Determine a sliding interval and a sliding step size of a sliding window of a preset length; Controlling the sliding window to slide on the application data after the grayscale version is released according to the sliding interval and the sliding step size of the sliding window; The sliding window is used to collect and aggregate the application data covered by each movement of the sliding window to obtain the target monitoring index of the grayscale version collected by the sliding window each time.
3. The method according to claim 1, characterized in that The method of fitting the standard monitoring indicators corresponding to the target monitoring indicators collected each time by the sliding window using the historical application data of the historical version includes: Determining the indicator type of the target monitoring indicator; Determining, according to the indicator type, target historical application data corresponding to the grayscale version from the historical application data of the historical version; The target historical application data is collected and aggregated using the sliding window to obtain an initial monitoring indicator corresponding to the target monitoring indicator collected each time by the sliding window; The initial monitoring indicator is fitted using a pre-trained fitting model to obtain a standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time.
4. The method according to claim 3, characterized in that The step of determining, according to the indicator type, target historical application data corresponding to the grayscale version from the historical application data of the historical version includes: When the indicator type is a ratio type, all the historical application data of the historical version are determined as the target historical application data corresponding to the grayscale version; When the indicator type is data type, the number of grayscale users corresponding to the grayscale version and the number of historical users corresponding to the historical version are obtained; Determining a user ratio of the grayscale version to the historical version according to the grayscale user quantity and the historical user quantity; The historical application data of the historical version is scaled according to the user ratio to obtain the target historical application data corresponding to the grayscale version.
5. The method according to claim 3, characterized in that: The using the sliding window to collect and aggregate the target historical application data to obtain the initial monitoring indicator corresponding to the target monitoring indicator collected by the sliding window each time includes: Determining a sliding interval and a sliding step size of the sliding window; Controlling the sliding window to slide on the target historical application data according to the sliding interval and sliding step size of the sliding window; The sliding window is used to collect and aggregate the target historical application data covered by each movement of the sliding window to obtain the initial monitoring index corresponding to the target monitoring index collected by the sliding window each time.
6. The method according to claim 1, characterized in that The performing anomaly detection on the grayscale version according to the target monitoring indicator collected each time by each sliding window and the standard monitoring indicator corresponding to the target monitoring indicator includes: For the target monitoring indicator collected each time in each sliding window, determine the standard indicator range corresponding to the target monitoring indicator according to the standard monitoring indicator corresponding to the target monitoring indicator and a preset anomaly detection algorithm; Determine whether the target monitoring indicator collected by each sliding window each time is within the corresponding standard indicator range; A sliding window in which the target monitoring indicator is outside the corresponding standard indicator range is determined as an abnormal sliding window; A sliding window in which the target monitoring indicator is within the corresponding standard indicator range is determined as a non-abnormal sliding window; Anomaly detection is performed on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window.
7. The method according to claim 6, characterized in that The performing anomaly detection on the grayscale version according to the abnormal sliding window and the non-abnormal sliding window includes: Determining whether the grayscale version is abnormal according to the abnormal sliding window; When it is determined according to the abnormal sliding window that the gray version does not have an abnormality, stop using the abnormal sliding window to collect and aggregate the application data after the gray version is released; When it is determined that the grayscale version is abnormal according to the abnormal sliding window, updating the abnormal sliding window to a non-abnormal sliding window; Using the non-abnormal sliding window to collect and aggregate the application data after the gray version is released, to obtain new target monitoring indicators collected by the non-abnormal sliding window; For each of the non-abnormal sliding windows, the historical application data of the historical version is used to fit the new standard monitoring indicator corresponding to the new target monitoring indicator collected each time by the non-abnormal sliding window; According to the new target monitoring indicator collected by each of the non-abnormal sliding windows and the new standard monitoring indicator corresponding to the new target monitoring indicator, anomaly detection is performed on the grayscale version.
8. A grayscale version of anomaly detection device, characterized in that: The device comprises: An indicator determination module is used to collect and aggregate application data after the grayscale version is released using a sliding window of a preset length for any sliding window of a preset length, and obtain a target monitoring indicator of the grayscale version collected each time by each sliding window; An indicator fitting module, used to fit the standard monitoring indicator corresponding to the target monitoring indicator collected by the sliding window using the historical application data of the historical version, wherein the historical version includes all versions before the gray version is released; The anomaly detection module is used to perform anomaly detection on the grayscale version according to the target monitoring indicator collected by each sliding window each time and the standard monitoring indicator corresponding to the target monitoring indicator.
9. An electronic device, characterized in that: include: A processor and a memory, wherein the processor is used to execute a grayscale version of anomaly detection program stored in the memory to implement the grayscale version of anomaly detection method according to any one of claims 1 to 7.
10. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the grayscale version of the abnormality detection method according to any one of claims 1 to 7.
