A data center host equipment alarm baseline dynamic generation method and related equipment

Abstract

The present invention discloses a method for dynamically generating an alarm baseline for a host device in a data center and related equipment, which can obtain performance status prediction data corresponding to the host device within the operating cycle range and access user ratio data and historical alarm ratio data corresponding to each time segment in the operating cycle range; using the performance status prediction data, access user ratio data and historical alarm ratio data, through tolerance calculation of different business system types, the initial alarm baseline corresponding to the host device within the operating cycle range is obtained; the initial alarm baseline is smoothed and filtered, and the host device is alarm-monitored using the filtered alarm baseline. The present invention uses the performance status prediction data of the host device, combined with the access user ratio data and historical alarm ratio data corresponding to the host device in each time segment to obtain a dynamic alarm baseline, effectively performs personalized alarm monitoring on each host device, and can accurately improve the operation and maintenance efficiency of the host device.

Description

Technical Field

[0001] The present invention relates to the field of data center IT equipment operation and maintenance technology, and in particular to a method for dynamically generating an alarm baseline for a data center host device and related equipment. Background Art

[0002] Data centers are key new information infrastructure resources. The host equipment carries an enterprise's various business application systems, data, and platform software, so their stable and reliable operation is crucial.

[0003] In the traditional host operation and maintenance management model, the alarm baseline mainly relies on the experience of the operation and maintenance personnel and is determined by manually inputting experience values. Since it cannot adapt to the personalized and lean operation and maintenance needs of massive hosts, and does not consider the differences in business systems carried by different hosts and their business importance, the equipment alarm monitoring in the data center has poor adaptability and low accuracy.

[0004] Therefore, how to effectively use dynamic baseline technology to perform alarm monitoring on host devices in a data center has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for dynamically generating an alarm baseline of a data center host device and related devices to overcome the above problems or at least partially solve the above problems. The technical solution is as follows:

[0006] A method for dynamically generating an alarm baseline of a host device in a data center, comprising:

[0007] Obtaining performance status prediction data corresponding to the operating cycle range of the host device;

[0008] Obtaining access user ratio data and historical alarm ratio data corresponding to each time segment of the host device within the operating cycle range;

[0009] Using the performance status prediction data, the access user ratio data and the historical alarm ratio data, and calculating tolerances of different business system types, an initial alarm baseline corresponding to the host device within the operating cycle range is obtained;

[0010] The initial alarm baseline is smoothed and filtered to obtain a filtered alarm baseline, so as to use the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle.

[0011] Optionally, obtaining the performance status prediction data corresponding to the operating cycle range of the host device includes:

[0012] Determine the operating cycle range;

[0013] The operating cycle range is input into a performance prediction model pre-built for the host device to obtain performance status prediction data output by the performance prediction model, wherein the performance prediction model is a data prediction model based on a gradient boosting decision tree algorithm.

[0014] Optionally, obtaining the access user ratio data and historical alarm ratio data corresponding to each time segment of the host device within the operating cycle range includes:

[0015] Obtaining a historical average user value and a historical average alarm value corresponding to each time segment of the host device within the operating cycle range;

[0016] Obtaining a total historical user value and a total historical alarm value corresponding to the host device within the operating cycle range;

[0017] For any of the time segments: the ratio of the historical average user value corresponding to the time segment to the historical total user value is determined as the access user ratio data corresponding to the time segment; the ratio of the historical average alarm value corresponding to the time segment to the historical total alarm value is determined as the historical alarm ratio data corresponding to the time segment.

[0018] Optionally, the using the performance status prediction data, the access user ratio data, and the historical alarm ratio data to calculate tolerances of different business system types to obtain an initial alarm baseline corresponding to the host device within the operating cycle range includes:

[0019] For any of the time segments: using the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data, and calculating the tolerance of different business system types, obtaining the performance tolerance index value corresponding to the time segment;

[0020] According to the performance tolerance index value corresponding to each of the time segments, an initial alarm baseline corresponding to the host device within the operating cycle range is obtained.

[0021] Optionally, performing smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline includes:

[0022] The initial alarm baseline is smoothed and filtered using a Savitzky-Golay filtering algorithm to obtain a filtered alarm baseline.

[0023] Optionally, the filtered alarm baseline is composed of performance alarm indicator values ​​corresponding to each of the time segments within the operating cycle range, and using the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle range includes:

[0024] Obtaining an actual performance indicator value of the host device at any time point within the operating cycle range;

[0025] The actual performance indicator value and the performance alarm indicator value corresponding to the time segment at the time point in the filtered alarm baseline are used to determine the alarm level of the host device at the time point and control the alarm signal light to output alarm information corresponding to the alarm level.

[0026] Optionally, the using the actual performance indicator value and the performance alarm indicator value corresponding to the time segment at the time point in the filtered alarm baseline to determine the alarm level of the host device at the time point and controlling the alarm signal light to output alarm information corresponding to the alarm level includes:

[0027] At 0.8line target >line true In the case of, it is determined that the alarm level of the host device at this time point is no alarm, and the alarm signal light is controlled to light green;

[0028] and / or, at 0.8line target ≤line true ≤0.9line target In the case of a warning, it is determined that the alarm level of the host device at this time point is a general alarm, and the alarm signal light is controlled to light yellow;

[0029] and / or, at 0.9line target <line true ≤line target In the case of a critical alarm, the alarm level of the host device at that time point is determined to be a critical alarm, and the alarm signal light is controlled to light up in red;

[0030] And / or, in line true >line target In the case of an emergency, the alarm level of the host device at that time point is determined to be an emergency alarm, and the alarm signal light is controlled to light up red and flash;

[0031] Among them, line true is the actual performance indicator value, line target It is the performance alarm indicator value corresponding to the time segment where the time point is located in the filtered alarm baseline.

[0032] Optionally, the using the performance status prediction data, the access user ratio data corresponding to the time segment, and the historical alarm ratio data to obtain the performance tolerance index value corresponding to the time segment includes:

[0033] According to the formula:

[0034] line base =y pred +0.5w user k(cos(πR user )+1)+0.5w warning k(cos(πR warning )+1)

[0035] Calculate the performance tolerance index value corresponding to the time segment, where 0≤line base ≤100,w user +w warning =1, line base is the performance tolerance index value corresponding to the time segment, y pred is the performance status prediction data, and w user Preset user proportion weight, w warning is the preset alarm ratio weight, k is the preset floating range parameter, R user is the access user ratio data, R warning It is the historical alarm ratio data.

[0036] A data center host equipment alarm baseline dynamic generation device comprises: a performance state prediction data acquisition unit, a ratio data acquisition unit, an initial alarm baseline acquisition unit and a filtered alarm baseline acquisition unit.

[0037] The performance status prediction data obtaining unit is used to obtain the performance status prediction data corresponding to the host device within the operating cycle range;

[0038] The proportion data obtaining unit is used to obtain the access user proportion data and the historical alarm proportion data corresponding to each time segment of the host device within the operation cycle range;

[0039] The initial alarm baseline obtaining unit is used to obtain the initial alarm baseline corresponding to the host device within the operating cycle range by using the performance status prediction data, the access user ratio data and the historical alarm ratio data through tolerance calculation of different business system types;

[0040] The filtered alarm baseline obtaining unit is used to perform smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline, so as to use the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle.

[0041] A computer-readable storage medium having a program stored thereon, characterized in that when the program is executed by a processor, the method for dynamically generating an alarm baseline for a data center host device described in any one of the above items is implemented.

[0042] An electronic device, comprising at least one processor, and at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other via the bus; and the processor is used to call program instructions in the memory to execute any of the above-mentioned methods for dynamically generating alarm baselines for data center host equipment.

[0043] By means of the above technical scheme, the present invention provides a method for dynamically generating an alarm baseline for a data center host device and related equipment, which can obtain the performance status prediction data corresponding to the host device within the operating cycle range, as well as the access user ratio data and historical alarm ratio data corresponding to each time segment in the operating cycle range; using the performance status prediction data, access user ratio data and historical alarm ratio data, through tolerance calculation of different business system types, the initial alarm baseline corresponding to the host device within the operating cycle range is obtained; the initial alarm baseline is smoothed and filtered to obtain a filtered alarm baseline, and the filtered alarm baseline is used to perform alarm monitoring on the host device within the operating cycle range. The present invention uses the performance status prediction data of the host device, combined with the access user ratio data and historical alarm ratio data corresponding to the host device in each time segment, to obtain a dynamic alarm baseline, effectively performing personalized alarm monitoring on each host device, and can accurately improve the operation and maintenance efficiency of the host device.

[0044] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0046] Figure 1 A schematic diagram showing a flow chart of an implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0047] Figure 2A flow chart showing another implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of the construction process of the performance prediction model provided by an embodiment of the present invention is shown;

[0049] Figure 4 A flow chart showing another implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0050] Figure 5 A flow chart showing another implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0051] Figure 6 A flow chart showing another implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0052] Figure 7 A schematic diagram showing a structure of a device for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention;

[0053] Figure 8 A schematic structural diagram of an electronic device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0054] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0055] In the traditional host operation and maintenance management mode, the operation and maintenance personnel mainly rely on querying the historical status curve of the host device operation to manually set the alarm threshold. The alarm threshold is theoretically unchanged unless it is manually adjusted. Therefore, the alarm monitoring method based on manually set alarm thresholds has poor adaptability and high labor costs. In addition, since the setting of the alarm threshold is based on manual experience, the accuracy is low. In order to adapt to the trend of intelligent operation and maintenance of host devices and improve the effectiveness and accuracy of equipment alarm monitoring, an embodiment of the present invention provides a method for dynamically generating alarm baselines for host devices in a data center, and uses the dynamically generated alarm baselines to complete alarm monitoring of host devices.

[0056] like Figure 1As shown, a flowchart of an implementation of a method for dynamically generating an alarm baseline of a data center host device provided by an embodiment of the present invention is provided. The method for dynamically generating an alarm baseline of a data center host device may include:

[0057] S100, obtaining performance status prediction data corresponding to the host device within an operating cycle range.

[0058] The host device may be an IT device in a data center, and may be a physical or virtual computer that provides a logically independent service unit, including its attached storage and an operating system running on the device.

[0059] The operation cycle range may be a window time that requires alarm monitoring. For example, the operation cycle range may be one day, one week, or one month. The embodiment of the present invention may predict the performance status prediction data of the host device within the operation cycle range in advance, and use the performance status prediction data to perform alarm monitoring on the host device within the operation cycle range.

[0060] The performance status prediction data may include a predicted CPU usage rate of the host device within an operating cycle range.

[0061] Optional, based on Figure 1 The method shown, such as Figure 2 As shown, a flowchart of another implementation of a method for dynamically generating an alarm baseline of a host device in a data center provided by an embodiment of the present invention is shown, and step S100 may include:

[0062] S110, determining the operation cycle range.

[0063] S120, inputting the operation cycle range into a performance prediction model pre-built for the host device to obtain performance status prediction data output by the performance prediction model, wherein the performance prediction model is a data prediction model based on a gradient boosting decision tree algorithm.

[0064] The operation cycle range may include multiple time points. In the embodiment of the present invention, multiple time points in the operation cycle range may be input into the performance prediction model, and the performance prediction model may predict the performance of the host device at each time point, and obtain the performance prediction values ​​corresponding to each time point, thereby obtaining the performance status prediction data of the host device in the operation cycle range.

[0065] Optional, such as Figure 3 As shown, the process of constructing the performance prediction model provided by the embodiment of the present invention may include:

[0066] A100: Obtain original historical statistical data of the host device, wherein the original historical statistical data may include historical performance data, historical alarm data, and historical user data.

[0067] Optionally, an embodiment of the present invention can count the CPU usage of the host device once every two minutes as historical performance data, and count the number of alarms and the number of users in each operating cycle of the previous month according to the window time of different operating cycles every day, which correspond to the historical alarm data and historical user data of the operating cycle.

[0068] The embodiment of the present invention can perform statistical processing on the original historical statistical data within a certain window time through periodic characteristic items such as weeks, months, and years and characteristic values ​​of the carried service system level according to a certain window time.

[0069] A200. Preprocess the original historical statistical data, remove abnormal data in the original historical performance data, and obtain the historical performance data to be trained.

[0070] Among them, abnormal data can be data that affects normal CPU usage when the host device finds faults during operation and maintenance, holidays, and major events. The embodiment of the present invention can remove abnormal data from the original historical performance data and retain the historical performance data to be trained that can reflect the law of change. The training data set and the test data set are divided according to the ratio of 8:2. Optionally, the embodiment of the present invention can also supplement the missing performance data.

[0071] A300, input the historical performance data in the historical performance data to be trained into the gradient boosting decision tree algorithm in the form of key-value pairs of time point-performance data for training and testing to obtain a performance prediction model.

[0072] A decision tree is a decision analysis method that uses a decision tree to construct a decision tree to find the probability that the expected value of the net present value is greater than or equal to zero, evaluate project risks, and judge its feasibility based on the known probabilities of various situations. It is a graphical method that intuitively uses probability analysis. A decision tree is a tree structure in which each internal node represents a test on an attribute, each branch represents a test output, and each leaf node represents a category. Since this decision branch is drawn as a graph that looks like a branch of a tree, it is called a decision tree. In machine learning, a decision tree is a prediction model that represents a mapping relationship between object attributes and object values. The embodiment of the present invention predicts performance data based on a ladder gradient boosting decision tree algorithm, adds a regularization term to the objective function, and each round of iteration is regularized at the mathematical level, greatly improving the effect and operability.

[0073] Specifically, assume that k is the number of decision trees, f k is the output result of the decision tree numbered k, and the prediction performance index is The loss function is in, T represents the number of leaves, w represents the weight of the leaf, and

[0074] Then, the loss function can be expressed as The Taylor expansion of the loss function is:

[0075] in, Removing the constant term gives Will I j ={i|q(x i )=j} is defined as the jth leaf point, that is: Then take the derivative of the above formula and set the derivative result equal to 0, we can get Put w j Substituting into the objective function, we get

[0076] It is understandable that the embodiment of the present invention can save the trained performance prediction model to predict the performance status of the host device in any future operating cycle range. For example: the embodiment of the present invention can predict the performance of the host device in the next 5 months and obtain the performance status prediction data for these 5 months. The operating cycle range is a window time in these 5 months. When the operating cycle range arrives, the performance status prediction data corresponding to the operating cycle range can be obtained from the performance status prediction data of these 5 months.

[0077] The embodiment of the present invention performs probability distribution calculation based on the real, minute-level historical statistical data of the host device, and adopts a performance prediction algorithm based on the gradient boosting decision tree to predict the minute-level performance indicator value of the device in any future operating cycle in units of days, so as to accurately predict the performance status prediction data of the host device.

[0078] S200: Obtain access user ratio data and historical alarm ratio data corresponding to each time segment within the operating cycle of the host device.

[0079] The time segment is a time period divided in the operation cycle range, for example, assuming that the operation cycle range is 24 hours in a day, then 0:00 to 7:00 can be a time segment, 7:00 to 8:00 can be a time segment, and so on. Preferably, the time segment in the embodiment of the present invention can be 8:00 to 18:00 every day.

[0080] The embodiment of the present invention can obtain access user ratio data and historical alarm ratio data respectively according to historical alarm data and historical user data.

[0081] Optional, based on Figure 1 The method shown, such as Figure 4As shown, a flow chart of another implementation of a method for dynamically generating an alarm baseline of a host device in a data center provided by an embodiment of the present invention is shown, and step S200 may include:

[0082] S210: Obtain the historical average user value and the historical average alarm value corresponding to each time segment in the operating cycle of the host device.

[0083] In the embodiment of the present invention, for any time segment, the following can be achieved: from historical user data, the amount of users in a first number of historical time segments corresponding to the time segment is obtained, and the ratio of the amount of users to the first number is determined as the historical average user value corresponding to the time segment. From historical alarm data, the amount of alarms in a second number of historical time segments corresponding to the time segment is obtained, and the ratio of the amount of alarms to the second number is determined as the historical average alarm value corresponding to the time segment.

[0084] S220: Obtain the total historical user value and the total historical alarm value corresponding to the host device within the operating cycle range.

[0085] The embodiment of the present invention can determine the sum of the user quantities in each historical operating cycle corresponding to the operating cycle range in the historical user data as the historical total user value corresponding to the operating cycle range, and determine the sum of the alarm quantities in each historical operating cycle corresponding to the operating cycle range in the historical alarm data as the historical total alarm value corresponding to the operating cycle range.

[0086] S230. For any time segment: the ratio of the historical average user value corresponding to the time segment to the historical total user value is determined as the access user ratio data corresponding to the time segment; the ratio of the historical average alarm value corresponding to the time segment to the historical total alarm value is determined as the historical alarm ratio data corresponding to the time segment.

[0087] Specifically, assuming n user is the historical average user value of the host device corresponding to this time segment, n warning N is the historical average alarm value of the host device corresponding to the time segment. user is the total number of historical users, N warning is the total historical alarm value, then the corresponding access user ratio data for this time segment is The historical alarm ratio data corresponding to this time segment is:

[0088] S300, using the performance status prediction data, the access user ratio data and the historical alarm ratio data, through tolerance calculation of different business system types, obtain the initial alarm baseline corresponding to the host device within the operating cycle range.

[0089] Optional, based on Figure 1 The method shown, such as Figure 5 As shown, a flowchart of another implementation of a method for dynamically generating an alarm baseline of a host device in a data center provided by an embodiment of the present invention is shown, and step S300 may include:

[0090] S310. For any time segment: using the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data, and calculating the tolerance of different business system types, obtain the performance tolerance index value corresponding to the time segment.

[0091] Optionally, the embodiment of the present invention may be based on the formula:

[0092] line base =y pred +0.5w user k(cos(πR user )+1)+0.5w warning k(cos(πR warning )+1)

[0093] Calculate the performance tolerance index value corresponding to the time segment, where 0≤line base ≤100,w user +w warning =1, line base is the performance tolerance index value corresponding to the time segment, y pred is the performance status prediction data, and w user Preset user proportion weight, w warning is the preset alarm ratio weight, k is the preset floating range parameter, R user To access user ratio data, R warning This is the historical alarm ratio data.

[0094] Among them, k can be set according to the alarm tolerance level of the business system where the host device is located. The embodiment of the present invention can divide the business system types into a first-class business system, a second-class business system and a third-class business system. Among them, a first-class business system may include an important production information system or a resource management information system in an enterprise. For example: an enterprise e-commerce mall and 95598 core business. A second-class system may include an enterprise resource management system and an intranet and extranet mail system. A third-class business system is a system responsible for other common businesses.

[0095] k represents the maximum range in which the performance tolerance index value can dynamically float. For a type of business system, there are many users and the usage is complex. In order to prevent failures or to handle failures as quickly as possible, assuming that the CPU usage rate of a device is 60% under normal use, the k value can be set to 25% for the devices under this business system, which means that an emergency alarm is required when the index reaches a maximum of 85%. For the third type of business system, the normal CPU usage rate may be only 50%, and the k value can be set to 35%. The performance tolerance index value refers to the acceptable degree of host operation risk during the operation and maintenance management of the host equipment in the data center. It is the tolerable limit of risks that occur during the operation of related equipment, which is set based on risk preference. The larger the performance tolerance index value, the stronger the ability of the equipment to withstand operation risks.

[0096] S320. Obtain an initial alarm baseline corresponding to the host device within the operating cycle range according to the performance tolerance index value corresponding to each time segment.

[0097] The embodiment of the present invention calculates the performance tolerance of the host device by adopting multi-dimensional constraints of the device as calculation parameters, divides the time into segments within the operating cycle range of the device, and takes into account the load conditions of different types of business systems in which the host device is located. The maximum range in which the performance tolerance index value can dynamically float is dynamically adjusted, so as to obtain a more accurate initial alarm baseline.

[0098] The embodiment of the present invention can accurately measure the complex alarm situation of device usage that continuously changes over time and under user usage by combining the historical alarm number and the historical user number in each time segment to determine the initial alarm baseline.

[0099] S400: Perform smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline, and use the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle.

[0100] Among them, the alarm baseline is the baseline of data changes, which is used to describe the normal fluctuation range of an indicator.

[0101] Optional, based on Figure 1 The method shown, such as Figure 6 As shown, a flowchart of another implementation of a method for dynamically generating an alarm baseline for a host device in a data center provided by an embodiment of the present invention is shown, and step S400 may include:

[0102] S410 , using a Savitzky-Golay filtering algorithm to perform smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline.

[0103] Savitzky-Golay filter algorithm, also known as Savitzky-Golay filter and SG filter. Savitzky-Golay filter is widely used in data stream smoothing and denoising. It is a filtering method based on local polynomial least squares fitting in the time domain. The biggest feature of this filter is that it can ensure that the shape and width of the signal remain unchanged while filtering out noise.

[0104] Specifically, the embodiment of the present invention can use the Savitzky-Golay filtering algorithm to perform smoothing filtering on the initial alarm baseline, perform polynomial least squares fitting on the data in the moving window, and calculate the weighted average sum of the center point in the window with respect to its surrounding points.

[0105] Specifically, the embodiment of the present invention can be based on the formula:

[0106]

[0107] Calculate the weighted average sum, where X i is the performance tolerance index value in the initial alarm baseline before smoothing, is the performance alarm index value in the alarm baseline after smoothing and filtering, that is, the weighted average sum. j is the weight factor in the moving window smoothing, and the length of the moving window is 2r+1.

[0108] The embodiment of the present invention performs smooth filtering on the sharp values ​​in the initial alarm baseline through the Savitzky-Golay filtering algorithm, and can obtain an effective and accurate filtered alarm baseline, so that the embodiment of the present invention can run smoothly in the device, thereby accurately determining the triggering conditions of subsequent alarm levels, thereby reducing the number of accidental false alarms.

[0109] Optionally, the filtered alarm baseline is composed of performance alarm index values ​​corresponding to each time segment in the operating cycle range. It is understandable that after the initial alarm baseline is smoothed and filtered to obtain the filtered alarm baseline, the performance tolerance index values ​​corresponding to each time segment on the initial alarm baseline are converted into performance alarm index values.

[0110] Optionally, the embodiment of the present invention can obtain the actual performance index value of the host device at any time point within the operating cycle range. Using the actual performance index value and the performance alarm index value corresponding to the time segment at the time point in the filtered alarm baseline, the alarm level of the host device at the time point is determined and the alarm signal light is controlled to output the alarm information corresponding to the alarm level.

[0111] The embodiment of the present invention can obtain the actual performance index value of the host device at any time point within the operating cycle in real time. It is understandable that the actual performance index value may include the CPU usage rate of the host device at that time point.

[0112] The embodiment of the present invention may compare the actual performance indicator value with the performance alarm indicator value corresponding to the time segment at the time point, and determine the alarm level of the host device at the time point according to the comparison result.

[0113] The embodiment of the present invention can pre-classify the alarm level into general alarm, important alarm and emergency alarm. The embodiment of the present invention can determine the alarm triggering condition for each alarm level based on the performance alarm index value corresponding to the time segment where the time point is located. For example: Assume that the performance alarm index value corresponding to the time segment where the time point is located is line target The actual performance index value of the host device at this time point is line true , then the alarm trigger condition for general alarms can be 0.8line target ≤line true ≤0.9line target , the trigger condition of the important alarm can be 0.9line target <line true ≤line target , the trigger condition of the emergency alarm can be line true >line target .

[0114] The embodiment of the present invention can track the alarm process and the main KPI indicator information causing the alarm in real time according to the actual operating status of the host device, realize hierarchical alarm of the host device status, and improve the operation and maintenance efficiency of the equipment.

[0115] Optionally, the embodiment of the present invention may be in line true The actual performance index value and line target When it is the performance alarm index value corresponding to the time segment of the time point in the filtered alarm baseline, it can be in 0.8line target >line true In the case of , the host device determines that the alarm level at that time point is no alarm, controls the alarm signal light to light green, and / or, at 0.8line target ≤line true ≤0.9line target In the case of a general alarm, the alarm level of the host device at that time point is determined to be a general alarm, and the alarm signal light is controlled to light yellow, and / or at 0.9line target <line true ≤line targetIn the case of a major alarm, the alarm level of the host device at that time point is determined to be a major alarm, and the alarm signal light is controlled to light red, and / or, in the line true >line target In the case of an emergency, the alarm level of the host device at that time point is determined to be an emergency alarm, and the alarm signal light is controlled to light up red and flash.

[0116] It should be noted that in actual situations, the actual performance indicator value may be positive or negative, so when compared, the actual performance indicator value and the performance alarm indicator value have the same sign.

[0117] The embodiment of the present invention can provide an early warning of the operating status of the host device by controlling the alarm signal light to output alarm information corresponding to the alarm level, making it convenient for technicians to perform corresponding maintenance on the host device according to the alarm level.

[0118] The present invention provides a method for dynamically generating an alarm baseline for a data center host device, which can obtain the performance status prediction data corresponding to the host device within the operating cycle range, as well as the access user ratio data and historical alarm ratio data corresponding to each time segment in the operating cycle range; using the performance status prediction data, access user ratio data and historical alarm ratio data, through tolerance calculation of different business system types, the initial alarm baseline corresponding to the host device within the operating cycle range is obtained; the initial alarm baseline is smoothed and filtered to obtain a filtered alarm baseline, and the filtered alarm baseline is used to perform alarm monitoring on the host device within the operating cycle range. The present invention uses the performance status prediction data of the host device, combined with the access user ratio data and historical alarm ratio data corresponding to the host device in each time segment to obtain a dynamic alarm baseline, effectively performs personalized alarm monitoring on each host device, and can accurately improve the operation and maintenance efficiency of the host device.

[0119] Although operations are depicted in a particular order, this should not be understood as requiring that the operations be performed in the particular order shown or in a sequential order.Multitasking and parallel processing may be advantageous under certain circumstances.

[0120] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0121] Corresponding to the above method embodiment, the embodiment of the present invention also provides a data center host equipment alarm baseline dynamic generation device, the structure of which is as follows: Figure 7As shown, it may include: a performance status prediction data obtaining unit 100 , a ratio data obtaining unit 200 , an initial alarm baseline obtaining unit 300 , and a filtered alarm baseline obtaining unit 400 .

[0122] The performance status prediction data obtaining unit 100 is used to obtain the performance status prediction data of the host device within the operating cycle range.

[0123] Optionally, the performance status prediction data obtaining unit 100 is specifically used to determine the operating cycle range; the operating cycle range is input into a performance prediction model pre-built for the host device to obtain performance status prediction data output by the performance prediction model, wherein the performance prediction model is a data prediction model based on a gradient boosting decision tree algorithm.

[0124] The ratio data obtaining unit 200 is used to obtain the access user ratio data and the historical alarm ratio data corresponding to each time segment in the operation cycle range of the host device.

[0125] Optionally, the proportion data acquisition unit 200 is specifically used to obtain the historical average user value and the historical average alarm value corresponding to each time segment in the operating cycle range of the host device; obtain the historical total user value and the historical total alarm value corresponding to the host device in the operating cycle range; for any time segment: the ratio of the historical average user value corresponding to the time segment to the historical total user value is determined as the access user proportion data corresponding to the time segment; the ratio of the historical average alarm value corresponding to the time segment to the historical total alarm value is determined as the historical alarm proportion data corresponding to the time segment.

[0126] The initial alarm baseline obtaining unit 300 is used to obtain the initial alarm baseline corresponding to the host device in the operating cycle range by using the performance status prediction data, the access user ratio data and the historical alarm ratio data, through the tolerance calculation of different business system types.

[0127] Optionally, the initial alarm baseline acquisition unit 300 is specifically used to: for any time segment: utilize the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data, and calculate the tolerance of different business system types to obtain the performance tolerance index value corresponding to the time segment; and obtain the initial alarm baseline corresponding to the host device in the operating cycle range according to the performance tolerance index value corresponding to each time segment.

[0128] Optionally, the embodiment of the present invention may be based on the formula:

[0129] line base =y pred +0.5w user k(cos(πR user)+1)+0.5w warning k(cos(πR warning )+1)

[0130] Calculate the performance tolerance index value corresponding to the time segment, where 0≤line base ≤100,w user +w warning =1, line base is the performance tolerance index value corresponding to the time segment, y pred is the performance status prediction data, and w user Preset user proportion weight, w warning is the preset alarm ratio weight, k is the preset floating range parameter, R user To access user ratio data, R warning This is the historical alarm ratio data.

[0131] The filtered alarm baseline obtaining unit 400 is used to perform smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline, so as to use the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle.

[0132] Optionally, the filtered alarm baseline obtaining unit 400 is specifically configured to perform smoothing filtering on the initial alarm baseline using a Savitzky-Golay filtering algorithm to obtain a filtered alarm baseline.

[0133] Optionally, the filtered alarm baseline is composed of performance alarm indicator values ​​corresponding to each time segment within the operating cycle range.

[0134] Optionally, an embodiment of the present invention can obtain the actual performance indicator value of the host device at any time point within the operating cycle range; use the actual performance indicator value and the performance alarm indicator value corresponding to the time segment of the time point in the filtered alarm baseline to determine the alarm level of the host device at that time point and control the alarm signal light to output alarm information corresponding to the alarm level.

[0135] Optionally, the embodiment of the present invention may be in line true The actual performance index value and line target When it is the performance alarm index value corresponding to the time segment of the time point in the filtered alarm baseline, it can be in 0.8line target >line true In the case of , the host device determines that the alarm level at that time point is no alarm, controls the alarm signal light to light green, and / or, at 0.8line target ≤line true ≤0.9line targetIn the case of a general alarm, the alarm level of the host device at that time point is determined to be a general alarm, and the alarm signal light is controlled to light yellow, and / or at 0.9line target <line true ≤line target In the case of a major alarm, the alarm level of the host device at that time point is determined to be a major alarm, and the alarm signal light is controlled to light red, and / or, in the line true >line target In the case of an emergency, the alarm level of the host device at that time point is determined to be an emergency alarm, and the alarm signal light is controlled to light up red and flash.

[0136] The present invention provides a data center host device alarm baseline dynamic generation device, which can obtain the performance status prediction data corresponding to the host device within the operating cycle range and the access user ratio data and historical alarm ratio data corresponding to each time segment in the operating cycle range; using the performance status prediction data, access user ratio data and historical alarm ratio data, through tolerance calculation of different business system types, to obtain the initial alarm baseline corresponding to the host device within the operating cycle range; smoothing and filtering the initial alarm baseline to obtain the filtered alarm baseline, and using the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle range. The present invention uses the performance status prediction data of the host device, combined with the access user ratio data and historical alarm ratio data corresponding to the host device in each time segment to obtain a dynamic alarm baseline, effectively performs personalized alarm monitoring on each host device, and can accurately improve the operation and maintenance efficiency of the host device.

[0137] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0138] The data center host equipment alarm baseline dynamic generation device includes a processor and a memory. The above-mentioned performance status prediction data acquisition unit 100, proportion data acquisition unit 200, initial alarm baseline acquisition unit 300 and filtered alarm baseline acquisition unit 400 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0139] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set. By adjusting the kernel parameters to use the performance status prediction data of the device, combined with the corresponding access user ratio data and historical alarm ratio data of each time segment, a dynamic alarm baseline can be obtained. Based on the alarm baseline, personalized alarm monitoring can be effectively performed on different devices, thereby improving the efficiency of equipment operation and maintenance.

[0140] An embodiment of the present invention provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the method for dynamically generating an alarm baseline for a host device in a data center is implemented.

[0141] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes the method for dynamically generating an alarm baseline for a data center host device when running.

[0142] like Figure 8 As shown, an embodiment of the present invention provides an electronic device 1000, which includes at least one processor 1001, at least one memory 1002 and a bus 1003 connected to the processor 1001; wherein the processor 1001 and the memory 1002 communicate with each other through the bus 1003; the processor 1001 is used to call the program instructions in the memory 1002 to execute the above-mentioned data center host device alarm baseline dynamic generation method. The electronic device in this article can be a server, PC, PAD, mobile phone, etc.

[0143] The present invention also provides a computer program product, which, when executed on an electronic device, is suitable for executing a program that initializes the steps of a method for dynamically generating an alarm baseline for a data center host device.

[0144] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatuses, electronic devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable device to generate a machine, so that the instructions executed by the processor of the computer or other programmable device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0145] In a typical configuration, an electronic device includes one or more processors (CPU), a memory, and a bus. The electronic device may also include an input / output interface, a network interface, and the like.

[0146] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.

[0147] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0148] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0149] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0150] The above are only embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for dynamically generating an alarm baseline for a host device in a data center, characterized in that: include: Obtaining performance status prediction data corresponding to the operating cycle range of the host device; Obtaining access user ratio data and historical alarm ratio data corresponding to each time segment of the host device within the operating cycle range; Using the performance status prediction data, the access user ratio data and the historical alarm ratio data, and calculating tolerances of different business system types, an initial alarm baseline corresponding to the host device within the operating cycle range is obtained; Performing smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline, and using the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle; Wherein, the using the performance status prediction data, the access user ratio data and the historical alarm ratio data, through tolerance calculation of different business system types, to obtain the initial alarm baseline corresponding to the host device within the operating cycle range, includes: For any of the time segments: using the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data, and calculating the tolerance of different business system types, obtaining the performance tolerance index value corresponding to the time segment; According to the performance tolerance index value corresponding to each of the time segments, obtaining an initial alarm baseline corresponding to the host device within the operating cycle range; Wherein, using the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data to obtain the performance tolerance index value corresponding to the time segment, includes: According to the formula: linebase=ypred+0.5wuserk(cos(πRuser)+1)+0.5wwarningk(cos(πRwarning)+1) calculates the performance tolerance index value corresponding to the time segment, where 0≤line base ≤100,w user +w warning =1, line base is the performance tolerance index value corresponding to the time segment, y pred is the performance status prediction data, w user is the preset user proportion weight, w warning is the preset alarm ratio weight, k is the preset floating range parameter, R user is the access user ratio data, R warning It is the historical alarm ratio data.

2. The method according to claim 1, characterized in that The obtaining of the performance status prediction data corresponding to the operating cycle range of the host device includes: Determine the operating cycle range; The operating cycle range is input into a performance prediction model pre-built for the host device to obtain performance status prediction data output by the performance prediction model, wherein the performance prediction model is a data prediction model based on a gradient boosting decision tree algorithm.

3. The method according to claim 1, characterized in that The obtaining of the access user ratio data and the historical alarm ratio data corresponding to each time segment of the host device within the operation cycle range includes: Obtaining a historical average user value and a historical average alarm value corresponding to each time segment of the host device within the operating cycle range; Obtaining a total historical user value and a total historical alarm value corresponding to the host device within the operating cycle range; For any of the time segments: the ratio of the historical average user value corresponding to the time segment to the historical total user value is determined as the access user ratio data corresponding to the time segment; the ratio of the historical average alarm value corresponding to the time segment to the historical total alarm value is determined as the historical alarm ratio data corresponding to the time segment.

4. The method according to claim 1, characterized in that: The step of performing smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline includes: The initial alarm baseline is smoothed and filtered using a Savitzky-Golay filtering algorithm to obtain a filtered alarm baseline.

5. The method according to claim 1, characterized in that The filtered alarm baseline is composed of performance alarm index values ​​corresponding to each of the time segments within the operating cycle range, and the use of the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle range includes: Obtaining an actual performance indicator value of the host device at any time point within the operating cycle range; The actual performance indicator value and the performance alarm indicator value corresponding to the time segment at the time point in the filtered alarm baseline are used to determine the alarm level of the host device at the time point and control the alarm signal light to output alarm information corresponding to the alarm level.

6. The method according to claim 5, characterized in that The using the actual performance indicator value and the performance alarm indicator value corresponding to the time segment at the time point in the filtered alarm baseline to determine the alarm level of the host device at the time point and controlling the alarm signal light to output alarm information corresponding to the alarm level includes: At 0.8line target >line true In the case of, it is determined that the alarm level of the host device at this time point is no alarm, and the alarm signal light is controlled to light green; and / or, at 0.8line target ≤line true ≤0.9line target In the case of a warning, it is determined that the alarm level of the host device at this time point is a general alarm, and the alarm signal light is controlled to light yellow; and / or, at 0.9line target <line true ≤line target In the case of a critical alarm, the alarm level of the host device at that time point is determined to be a critical alarm, and the alarm signal light is controlled to light up in red; And / or, in line true >line target In the case of an emergency, the alarm level of the host device at that time point is determined to be an emergency alarm, and the alarm signal light is controlled to light up red and flash; Among them, line true is the actual performance indicator value, line target It is the performance alarm indicator value corresponding to the time segment where the time point is located in the filtered alarm baseline.

7. A data center host equipment alarm baseline dynamic generation device, characterized in that: include: performance status prediction data acquisition unit, ratio data acquisition unit, initial alarm baseline acquisition unit and filtered alarm baseline acquisition unit, The performance status prediction data obtaining unit is used to obtain the performance status prediction data corresponding to the host device within the operating cycle range; The proportion data obtaining unit is used to obtain the access user proportion data and the historical alarm proportion data corresponding to each time segment of the host device within the operation cycle range; The initial alarm baseline obtaining unit is used to obtain the initial alarm baseline corresponding to the host device within the operating cycle range by using the performance status prediction data, the access user ratio data and the historical alarm ratio data, through tolerance calculations of different business system types, including: for any of the time segments: using the performance status prediction data, the access user ratio data corresponding to the time segment and the historical alarm ratio data, through tolerance calculations of different business system types, to obtain the performance tolerance index value corresponding to the time segment; according to the performance tolerance index value corresponding to each of the time segments, obtaining the initial alarm baseline corresponding to the host device within the operating cycle range; The filtered alarm baseline obtaining unit is used to perform smoothing filtering on the initial alarm baseline to obtain a filtered alarm baseline, so as to use the filtered alarm baseline to perform alarm monitoring on the host device within the operating cycle; The initial alarm baseline obtaining unit obtains the performance tolerance index value corresponding to the time segment by using the performance status prediction data, the access user ratio data corresponding to the time segment, and the historical alarm ratio data, including: According to the formula: linebase=ypred+0.5wuserk(cos(πRuser)+1)+0.5wwarningk(cos(πRwarning)+1) Calculate the performance tolerance index value corresponding to the time segment, where 0≤line base ≤100,w user +w warning =1, line base is the performance tolerance index value corresponding to the time segment, y pred is the performance status prediction data, w user is the preset user proportion weight, w warning is the preset alarm ratio weight, k is the preset floating range parameter, R user is the access user ratio data, R warning It is the historical alarm ratio data.

8. An electronic device, comprising at least one processor, and at least one memory and bus connected to the processor; wherein: The processor and the memory communicate with each other via the bus; The processor is used to call the program instructions in the memory to execute the method for dynamically generating an alarm baseline for a data center host device according to any one of claims 1 to 6.

Images