Time series data management method and device and computing equipment

Abstract

The invention discloses a time series data management method, which comprises the following steps that: a management device acquires a data set of time series data and stores the acquired data set of the time series data into a storage space; the management device aging the data set of the time series data in the storage space; wherein the acquisition speed of the data set of the time series data acquired by the management device is positively correlated with the available capacity of the storage space, and / or the aging speed of the data set of the time series data in the storage space is negatively correlated with the available capacity; the available capacity is the available capacity in the storage space for managing the data set of the time series data stored by the device. The method can effectively prevent the capacity of the storage space of the time series data from exceeding the limit.

Description

Technical Field

[0001] This application relates to the field of computer technologies, and particularly to a method and apparatus for managing time series data and a computing device. Background Art

[0002] Currently, computer systems such as storage systems have become extremely large, and it is necessary to monitor the health status of each module (software module, hardware module) in the computer system. In common health status monitoring methods, performance data generated during the operation of each module is continuously acquired and stored to obtain time series data, so that users can evaluate the health status of the computer system by viewing the performance data of each module.

[0003] Currently, the acquired performance data is stored in the following manner.

[0004] 1. Cache the acquired performance data in the system disk of the computer system. Since the performance data is time series data, caching the time series data in the system disk requires frequent writing to the system disk, which reduces the lifespan of the system disk. In addition, the capacity of the system disk is limited, and only a small amount of recently generated performance data can be cached, resulting in the inability to store the performance data for a long time.

[0005] 2. Store the acquired performance data in the data disk of the computer system. To control costs, the data disk used to store performance data has a capacity limit. Since there are many modules in the computer system and the amount of acquired performance data is large, it is easy to cause the data disk capacity to exceed the limit, and then the recently acquired performance data cannot be written. Summary of the Invention

[0006] A method and apparatus for managing time series data and a computing device are provided, which can effectively prevent the capacity of the storage space for time series data from exceeding the limit.

[0007] In a first aspect, a method for managing time series data is provided, where the time series data includes a data set generated in chronological order; the method includes: a management device acquires the data set of the time series data and stores the acquired data set of the time series data in a storage space; the management device ages the data set of the time series data in the storage space; where the acquisition speed of the data set of the time series data acquired by the management device is positively correlated with the available capacity of the storage space, and / or, the aging speed of the data set of the time series data in the storage space by the management device is negatively correlated with the available capacity; the available capacity is the available capacity of the storage space for the management device to store the data set of the time series data.

[0008] Among them, the time series data acquired by the management device may be the performance data of each module in the computer system. The health status of the computer system can be analyzed through these performance data.

[0009] The faster the management device acquires the dataset of time-series data, the faster the dataset is stored in the storage space, and the faster the available capacity of the storage space is consumed. On the contrary, the slower the management device acquires the dataset of time-series data, the larger the granularity of dataset acquisition, that is, the larger the time interval between adjacent datasets in the time-series data. This is not conducive to detailed analysis based on time-series data and reduces the accuracy of the analysis results. When the management device acquires the dataset of time-series data and stores the acquired dataset in the storage space, the acquisition speed of the management device for the dataset can be dynamically adjusted based on the dynamic change of the available capacity of the storage space. Among them, a large available capacity indicates that there is enough space to store the dataset. In this case, the acquisition speed is increased to acquire the dataset of time-series data with a finer granularity, so that more detailed analysis can be performed based on the time-series data. A small available capacity indicates that the space for storing the dataset is tight. In this case, the acquisition speed is reduced, and then the speed of writing the dataset into the storage space is reduced. In this way, the capacity overrun of the storage space can be avoided.

[0010] The faster the management device ages the dataset in the storage space, the more data is deleted from the storage space per unit time, and the more capacity can be released. However, the faster the management device ages the dataset, the shorter the storage time of the dataset, which is not conducive to comprehensive analysis based on time-series data. The aging speed of the management device for the dataset can be dynamically adjusted based on the dynamic change of the available capacity of the storage space. Among them, a large available capacity indicates that there is enough capacity to store data. In this case, the aging speed is reduced to extend the storage time of the dataset in the storage space, so that more comprehensive analysis can be performed based on the dataset in the storage space. A small available capacity indicates that the capacity in the storage space is tight and more capacity needs to be released, otherwise capacity overrun may occur. In this case, the aging speed is increased to quickly release the capacity and avoid capacity overrun of the storage space.

[0011] In summary, in the process of acquiring and storing time-series data, the acquisition strategy and / or aging strategy can be dynamically adjusted based on the available capacity of the storage space, so that the acquisition speed and / or aging speed of the time-series data match the available capacity, thereby maintaining the available capacity at a reasonable level and avoiding capacity overrun.

[0012] In a possible implementation, the management device acquires the dataset of time-series data, including: the management device performs data collection on the target object to obtain the dataset of time-series data; among them, the collection times of the data in different datasets of the time-series data are different.

[0013] In this implementation, the management device can acquire the dataset of time-series data by performing data collection on the target object. The acquired dataset of time-series data can be used to analyze the health status of the target object, etc.

[0014] The acquisition speed at which the management device obtains the data set is the acquisition speed for data collection of the target object. Among them, the acquisition speed and the acquisition granularity are negatively correlated. In other words, the available capacity and the acquisition granularity are negatively correlated. Thus, when the available capacity is large, data collection of the target object can be performed according to a small acquisition granularity, so as to collect a fine-grained data set (that is, the time interval between adjacent data sets is short). The finer the granularity of the data set in the time-series data, the more details the time-series data includes. Thus, the accuracy of analysis based on the time-series data can be improved. When the available capacity is small, data collection of the target object can be performed according to a large acquisition granularity, reducing the data sets collected per unit time, and thus reducing the data sets stored in the storage space per unit time, which can avoid excessive occupation of the storage space by the time-series data and further avoid exceeding the capacity of the storage space.

[0015] In a possible implementation manner, the acquisition speed includes the acquisition speed at which the management device performs data collection on the target object; the management device performing data collection on the target object includes: the management device performing data collection on the target object at a first acquisition speed; when the data set of the time-series data includes abnormal data, the management device performing data collection on the target object at a second acquisition speed; where the second acquisition speed is greater than the first acquisition speed.

[0016] The abnormal data can reflect the abnormality that occurs to the target object. If the data of the time-series data includes abnormal data, it indicates that the target object may have an abnormality. In this case, performing data collection on the target object at a faster acquisition speed, that is, performing data collection on the target object at a smaller acquisition granularity, can make the time-series data include more details, so that more detailed analysis can be performed based on the time-series data to more accurately identify the abnormality that occurs to the target object.

[0017] In a possible implementation manner, the management device ages the data sets of the time-series data in the storage space, including: the management device aging the data sets of the time-series data in the storage space at a first aging speed; when the data sets of the time-series data in the storage space include abnormal data, the management device aging the data sets of the time-series data in the storage space at a second aging speed; where the second aging speed is less than the first aging speed.

[0018] When the data set of the time-series data includes abnormal data, it indicates that the time-series data may reflect an abnormal situation and has great analysis value. Aging the data set including abnormal data at a small aging speed can enable the data set with great analysis value to be stored for a longer time, thus meeting the subsequent abnormal analysis requirements.

[0019] In a possible implementation, the time series data includes a key data set and a non-key data set; among them, the key data set is the data set in the time series data that reflects the change trend of the time series data, and the non-key data set is the data set in the time series data other than the key data set; the management device ages the data sets of the time series data in the storage space, including: aging the non-key data sets of the time series data in the storage space.

[0020] Among them, the key data set in the time series data can be obtained through the time series data downsampling algorithm, and then the non-key data set in the time series data can be obtained.

[0021] The key data set reflects the change trend of the time series data and has great analysis value. The non-key data set has little impact on the change trend of the time series data and has little analysis value. Aging the non-key data sets of the time series data while retaining the key data sets of the time series data can reduce the occupancy of the storage space by the time series data while retaining the data sets that can reflect the change trend of the time series data, and can meet the needs of subsequent related analyses.

[0022] In a possible implementation, the method includes: receiving a display instruction for the time series data; in response to the display instruction, supplementing the non-key data set for the time series data; and displaying the time series data based on the supplemented non-key data set and the data sets of the time series data stored in the storage space.

[0023] The aging operation on the time series data deletes the non-key data sets in the time series data. Although the non-key data sets have little impact on the change trend of the time series data, they can reflect the details of the time series data. When it is necessary to display the time series data, the non-key data sets of the time series data can be supplemented to supplement the non-key data sets deleted by the aging operation. Displaying the time series data based on the supplemented non-key data set and the data sets retained in the storage space (i.e., the key data sets) can display the time series data with richer details.

[0024] In a possible implementation, storing the obtained data sets of the time series data in the storage space includes: inputting the data sets of the time series data into a pre-trained encoding module so that the encoding module outputs the encoding results of the data sets of the time series data; and storing the encoding results in the storage space.

[0025] The encoding result is the encoding vector obtained by the encoding module extracting features from the data set. The encoding result of the data set can reflect the data set features, and the data volume is smaller than the data volume of the data set itself. Storing the encoding results of the data sets of the time series data in the storage space can not only store the features of the data sets, but also reduce the data storage volume in the storage space and save the available capacity of the storage space.

[0026] In a possible implementation, the encoding module corresponds to the decoding module, and the decoding module is used to decode the encoding result into a dataset of time-series data; the method includes: receiving a display instruction for the time-series data; in response to the display instruction, inputting the encoding result into the decoding module, so that the decoding module outputs a dataset of time-series data; and displaying the time-series data based on the dataset of time-series data output by the decoding module.

[0027] Among them, the encoding module and the decoding module belong to the same AI model (such as a generative model). The encoding module and the decoding module can be obtained by training the AI model.

[0028] When it is necessary to display the time-series data, through the decoding module, the encoding result is decoded into a dataset of time-series data, so as to obtain a dataset of time-series data, and then the time-series data can be displayed based on the dataset of time-series data. Thus, while reducing the amount of data stored in the storage space, the normal display of the time-series data is ensured.

[0029] In a second aspect, a time-series data management device is provided, where the time-series data includes a dataset generated in chronological order; the method includes: an acquisition module, configured to acquire a dataset of time-series data and store the acquired dataset of time-series data in a storage space; an aging module, configured to age the dataset of time-series data in the storage space; where the acquisition speed of the acquisition module for acquiring the dataset of time-series data is positively correlated with the available capacity of the storage space, and / or, the aging speed of the aging module for aging the dataset of time-series data in the storage space is negatively correlated with the available capacity; the available capacity is the available capacity in the storage space for the management device to store the dataset of time-series data.

[0030] In a possible implementation, the acquisition module is configured to: perform data acquisition on a target object to obtain a dataset of time-series data; where the acquisition times of the data in different datasets of the time-series data are different.

[0031] In a possible implementation, the acquisition speed includes the acquisition speed of the acquisition module for performing data acquisition on the target object; the acquisition module is configured to: perform data acquisition on the target object at a first acquisition speed; when the dataset of time-series data includes abnormal data, perform data acquisition on the target object at a second acquisition speed; where the second acquisition speed is greater than the first acquisition speed.

[0032] In a possible implementation, the aging module is configured to: age the dataset of time-series data in the storage space at a first aging speed; when the dataset of time-series data in the storage space includes abnormal data, age the dataset of time-series data in the storage space at a second aging speed; where the second aging speed is less than the first aging speed.

[0033] In a possible implementation, the time-series data includes a critical data set and a non-critical data set; wherein, the critical data set is the data set in the time-series data that reflects the change trend of the time-series data, and the non-critical data set is the data set in the time-series data other than the critical data set; the aging module is used to age the non-critical data set of the time-series data in the storage space.

[0034] In a possible implementation, the management device further includes: a first receiving module, configured to receive a display instruction for the time-series data; a supplement module, configured to supplement the non-critical data set for the time-series data in response to the display instruction; and a first display module, configured to display the time-series data based on the supplemented non-critical data set and the data set of the time-series data stored in the storage space.

[0035] In a possible implementation, the acquisition module is used to: input the data set of the time-series data into a pre-trained encoding module, so that the encoding module outputs an encoding result of the data set of the time-series data; and store the encoding result in the storage space.

[0036] In a possible implementation, the encoding module corresponds to a decoding module, and the decoding module is used to decode the encoding result into the data set of the time-series data; the management device further includes: a second receiving module, configured to receive a display instruction for the time-series data; an input module, configured to input the encoding result into the decoding module in response to the display instruction, so that the decoding module outputs the data set of the time-series data; and a second display module, configured to display the time-series data based on the data set of the time-series data output by the decoding module.

[0037] In a third aspect, a computing device is provided, including: a memory, configured to store an executable program; and a processor, configured to execute the method provided in the first aspect by running the executable program.

[0038] In a fourth aspect, a computer-readable storage medium is provided, including computer program instructions, and when the computer program instructions are executed by a computing device, the computing device executes the method provided in the first aspect.

[0039] In a fifth aspect, a computer program product including instructions is provided, and when the instructions are run by a computer device, the computer device executes the method provided in the first aspect.

[0040] Among them, the beneficial effects of the second aspect to the fifth aspect can be referred to the description of the beneficial effects of the first aspect above, and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic structural diagram of a system architecture provided by an embodiment of the present application;

[0042] Figure 2A structural schematic diagram of a system architecture provided by an embodiment of the present application;

[0043] Figure 3 A schematic diagram of an aging strategy provided by an embodiment of the present application;

[0044] Figure 4 A schematic diagram of a user interface provided by an embodiment of the present application;

[0045] Figure 5 A flowchart of a method for managing time-series data provided by an embodiment of the present application;

[0046] Figure 6 A schematic diagram of a method for managing time-series data provided by an embodiment of the present application;

[0047] Figure 7 A structural schematic diagram of a device for managing time-series data provided by an embodiment of the present application;

[0048] Figure 8 A structural schematic diagram of a computing device provided by an embodiment of the present application. Detailed implementation manners

[0049] The solutions provided by the embodiments of the present application will be described below in conjunction with the accompanying drawings. Among them, in the embodiments of the present application, "a plurality of" means two or more than two. "First", "second", etc. are only used to distinguish similar objects and do not necessarily need to describe a specific order or the number of objects.

[0050] To facilitate the understanding of the solutions provided by the embodiments of the present application, before introducing the solutions provided by the embodiments of the present application, some technical terms that may be involved in the embodiments of the present application will be introduced first.

[0051] Time-series data: Also known as streaming data, it refers to data generated in chronological order, that is, data continuously generated over time. Time-series data consists of data sets generated in chronological order. Among them, each data set can include one or more data, and the generation times of the data in different data sets are different. Specifically, at a time point or time period, one or more data can be generated, and the one or more data can be called a data set. Data sets generated at multiple adjacent time points or time periods in sequence form a segment of time-series data. For data acquisition, each data acquisition operation on the target object generates a data set, and multiple adjacent data acquisition operations in sequence generate multiple data sets, and the multiple data sets form a segment of time-series data. Time-series data often has characteristics such as a large amount of data, a short interval between data sets, and continuous arrival of data sets.

[0052] Data acquisition strategy: Abbreviated as acquisition strategy, it is a related rule for controlling data or data set acquisition behavior. The acquisition strategy can include acquisition speed, data source, priority level of the data source, etc.

[0053] Data source: It refers to the origin of the acquired data, that is, the data or dataset is obtained from the data source. Different data sources can have different priority levels, and the speeds of obtaining data or datasets from data sources with different priority levels are different. Among them, the priority level of the data source is positively correlated with the acquisition speed of the data source.

[0054] Acquisition speed: It refers to the speed of acquiring data.

[0055] Collection strategy: A data acquisition strategy, including collection speed, collection target object, priority level of the collection target object, etc. Among them, the collection target object can be simply referred to as the target object, which is a type of data source.

[0056] Collection granularity: It is the reciprocal of the execution frequency of the data collection operation and is the time interval of the data collection operation performed on the target object.

[0057] Collection speed: It is inversely proportional to the collection granularity. The smaller the collection granularity, the more frequent the data collection operation, and the greater the data collection speed. Among them, the priority level of the target object is positively correlated with the collection speed of the target object.

[0058] Aging: Also known as the aging operation, it refers to the control of the life cycle of the dataset in the time-series data, and can also be understood as the operation of deleting the dataset in the time-series data. Generally speaking, the storage duration of the dataset can be set, and when the storage duration reaches, the dataset is deleted. Or rather, the length of the life cycle of the dataset can be set, and when the life cycle of the dataset ends, the dataset is deleted.

[0059] Data aging strategy: Abbreviated as the aging strategy, it is a strategy used to age the time-series data. This strategy can age the time-series data by aging the dataset of the time-series data. The aging strategy includes aging object, aging speed, aging priority level of the aging object, etc.

[0060] Aging object: It is the time-series data or time-series data segment. Among them, the aging priority levels of different aging objects are different. The aging priority level of the aging object is positively correlated with the aging speed of the aging object.

[0061] Aging speed: For a period of time-series data, the aging speed can refer to the number of datasets deleted in one aging operation. The more datasets deleted, the faster the aging speed. The aging speed can also refer to the frequency of performing the aging operation. The higher the frequency, the faster the aging speed.

[0062] Time series data segment: It refers to a segment of time series data composed of multiple data sets obtained within a period of time. For example, a segment of time series data composed of multiple data sets obtained within one day, a segment of time series data composed of multiple data sets obtained within two days, and so on. Aging operations can be performed on the time series data segment to delete data sets in the time series data segment.

[0063] Storage space: The space used to store data. Among them, the storage space can be one or more disks, or one or more partitions in a disk. Exemplarily, the storage space can be implemented as a storage volume.

[0064] Available capacity: It refers to the remaining, unoccupied capacity or space in the storage space.

[0065] Key data set: It can also be an important data set, which refers to the data set in the time series data that can reflect the change trend of the time series data. The key data set in the time series data can be calculated through an algorithm. For example, the key data set in the time series data can be obtained through time series data downsampling algorithms such as time series data dynamic downsampling algorithm, perceptually important points (PIP) algorithm, frequency domain feature extraction algorithm, and time domain feature extraction algorithm. Among them, for the same segment of time series data, when the calculation parameters of the same algorithm are different, the number of key data sets obtained may be different. Taking the PIP algorithm as an example, this algorithm smooths the time series data, then calculates the weight of each data set in the time series data, and then selects the N data sets with the highest weights as important points. Among them, N is the calculation parameter of the PIP algorithm.

[0066] Non-key data set: It is the data set in the time series data except the key data set. Among them, in the embodiments of the present application, the aging operation may refer to deleting the non-key data sets in the time series data segment, and the adjustment of the aging speed can be achieved by adjusting the proportion of non-key data sets in the time series data segment. When the proportion of non-key data sets in the time series data set increases, the aging speed of the time series data segment also increases. Among them, by adjusting the calculation parameters of the time series data downsampling algorithm (such as N in the above-mentioned PIP algorithm), the proportion of non-key data sets in the time series data set is adjusted.

[0067] The embodiments of the present application provide a time series data management method. In this method, the management device obtains the data sets of the time series data according to the acquisition strategy, stores the obtained data sets of the time series data in the storage space, and ages the data sets of the time series data in the storage space according to the aging strategy. Among them, the management device can obtain the available capacity of the storage space used to store the data sets of the time series data, and adjust the acquisition strategy and / or the aging strategy based on the available capacity.

[0068] Among them, the acquisition strategy includes the acquisition speed. The acquisition speed is positively correlated with the available capacity, that is, the smaller the available capacity, the smaller the acquisition speed. The management device continues to acquire the dataset of the time series data according to the adjusted acquisition speed, and stores the acquired dataset of the time series data into the storage space. In this way, when the available capacity of the storage space is small, the speed at which the management device acquires the dataset of the time series data is slowed down, and thus the speed at which the dataset of the time series data is written into the storage space is slowed down, avoiding the over-limit of the storage space capacity.

[0069] The aging strategy includes the aging speed. The aging speed is negatively correlated with the available capacity, that is, the smaller the available capacity, the greater the aging speed. The management device continues to age the dataset of the time series data in the storage space according to the adjusted aging strategy. In this way, when the available capacity of the storage space is small, the aging of the dataset of the time series data in the storage space is accelerated, avoiding the over-limit of the storage space capacity.

[0070] Next, the time series data management method provided by the embodiments of the present application will be introduced.

[0071] Figure 1 A system architecture is shown. This system architecture can be used to implement the time series data management method provided by the embodiments of the present application. As Figure 1 shown, the system architecture includes a management device 100 and a storage space 200. The management device 100 can be any device, equipment, cluster or platform with data processing and communication functions. The storage space 200 is a storage space for persistently storing data. The storage space 200 can be mounted to the management device 100, so that the management device 100 can store data into the storage space 200 and acquire the data in the storage space 200. The storage space 200 can be located locally to the management device 100 or remotely from the management device 100. The storage space 200 can be composed of one or more disks, or one or more partitions in a disk. In some embodiments, the storage space 200 can specifically be a storage volume, such as a performance management system (PMS) volume.

[0072] In some embodiments, the storage space 200 can be set by a user (such as an operation and maintenance personnel). In some embodiments, the management device 100 can select one or more disks from a storage pool and use the selected disks as the storage space 200. In some embodiments, the management device 100 can select one or more partitions from a storage pool and use the selected partitions as the storage space 200.

[0073] The management device 100 includes an acquisition module 110. The acquisition module 110 can acquire a data set of time-series data according to an acquisition strategy. Among them, the acquisition strategy can include multiple acquisition speeds, and different acquisition speeds correspond to different acquisition priority levels. The acquisition module 110 acquires a data set from the data source according to the acquisition speed corresponding to the priority level of the data source.

[0074] After the acquisition module 110 is started, it can acquire a data set of time-series data according to the initialized acquisition strategy. Among them, the initialized acquisition strategy includes a preset acquisition speed and a preset data source. Exemplarily, the preset data source has a preset acquisition priority. When there are multiple preset data sources, the preset acquisition priority levels of different data sources may be different or the same. The preset acquisition priority of the data source corresponds to the preset speed of the data source.

[0075] In some embodiments, the acquisition module 110 has a data acquisition function. The data set of time-series data acquired by the acquisition module 110 is specifically obtained by the acquisition module 110 performing data acquisition on a target object. For example, the acquisition module 110 can perform data acquisition on the target object A1 to obtain a data set of time-series data A11. For another example, the acquisition module 110 can perform data acquisition on the target object A2 to obtain a data set of time-series data A21. And so on.

[0076] The target object can be a hardware module in a computer system such as a storage system, such as a central processing unit (CPU), memory, network card, storage array, etc. The target object can also be a software module in a computer system, such as an operating system. Among them, a data acquisition module runs on the target object, and the data acquisition module can perform data acquisition operations on the target object. Exemplarily, the data acquisition module can be a process or a thread. For each data acquisition operation on the target object, a data set of time-series data is obtained. For consecutive multiple operations on the target object, multiple sequentially adjacent data sets of time-series data are obtained, and these multiple data sets form a segment of time-series data.

[0077] The acquisition strategy can specifically be an acquisition policy. The acquisition policy can include an acquisition granularity (i.e., the time interval between two adjacent data acquisition operations), the target object to be acquired, etc. When there are multiple target objects, the acquisition policy can include the priority level of the target object, etc. Among them, the acquisition granularity of the target object with a high priority level is smaller than that of the target object with a low priority level, that is, the acquisition speed of the target object with a high priority level is greater than that of the target object with a low priority level. Among them, after the acquisition module 110 is started, it performs data acquisition using the initialized acquisition policy. The initialized acquisition policy includes a preset acquisition granularity (such as 5 seconds), a preset target object, and the priority level of the target object, etc.

[0078] The acquisition module 110 stores the acquired dataset of timing data in the storage space 200.

[0079] In some embodiments, as Figure 1 shown, the acquisition module 110 may store the acquired dataset of timing data in the storage space 200.

[0080] In some embodiments, as Figure 2 shown, the acquisition module 110 may send the acquired dataset of timing data to the processing module 130, and the processing module 130 stores the dataset of timing data in the storage space 200.

[0081] In some embodiments, the acquisition module 110 or the processing module 130 may perform semantic processing on the dataset of timing data, and store the dataset after semantic processing in the storage space 200. Among them, semantic processing refers to processing the dataset into data that can directly reflect performance indicators. For example, the transmission delay of the data at the sending end and the receiving end is calculated through the sending time of the data sent by the sending end and the receiving time of the receiving end receiving the data.

[0082] In some embodiments, the acquisition module 110 or the processing module 130 may perform compression processing on the dataset of timing data. Among them, compression processing is also called refining or summarizing, which is an operation to reduce the amount of data. In one example, the dataset can be compressed through compression algorithms such as run-length encoding (RLE) algorithm and Huffman algorithm.

[0083] In some embodiments, an artificial intelligence (AI) model, such as a generative model, can be trained using the existing dataset of timing data. The model includes an encoding module and a decoding module. Among them, the encoding module is used to extract the features of the dataset of timing data to obtain an encoded vector. Among them, the encoded vector is also called the encoding result. The amount of data of the encoded vector of the dataset of timing data is smaller than the amount of data of the dataset of timing data. The decoding module can generate the dataset of timing data based on the encoded vector. The acquisition module 110 or the processing module 130 may input the acquired dataset of timing data into the encoding module, so that the encoding module outputs the encoding result of the dataset of timing data. Then, the acquisition module 110 or the processing module 130 stores the encoding result of the dataset of timing data in the storage space 200. When it is necessary to display the timing data, the encoding structure in the storage space 200 is input into the decoding module, so that the decoding module outputs the dataset of timing data. Furthermore, the timing data can be displayed based on the dataset of timing data.

[0084] The management device 100 further includes an aging module 120. The aging module 120 can age the data sets of the time-series data in the storage space 200 according to an aging policy. In some embodiments, the aging policy may include multiple aging speeds, where the magnitudes of different aging speeds are different, and different aging speeds correspond to different aging priority levels. The aging module 120 can age the aging object according to the aging speed corresponding to the aging priority level of the aging object.

[0085] After the aging module 120 is started, it can age the data sets of the time-series data in the storage space 200 according to the initialized aging policy. The initialized aging policy includes a preset aging speed, a preset aging object, a preset priority level of the aging object, the storage space write speed, etc. Among them, the preset priority levels of different aging objects can be different or the same. The preset aging object corresponds to the data source in the initialized acquisition policy, that is, the aging object is the time-series data acquired from the data source and stored in the storage space 200.

[0086] In some embodiments, the data sets of the time-series data in the storage space 200 can be aged based on a time-series data segment. The aging of the data sets in the time-series data segment specifically refers to deleting the non-critical data sets in the time-series data segment. Specifically, the critical data sets in the time-series data segment can be determined according to a time-series data downsampling algorithm such as the PIP algorithm, and the data sets other than the critical data sets are used as non-critical data sets. The aging operation on the time-series data segment is specifically to delete the non-critical data sets in the time-series data segment. In an example, refer to Figure 3 , it can be set that 10,000 data sets in the time-series data are acquired on October 23, and these 10,000 data sets form a time-series data segment. Assuming that 8,000 critical data sets and 2,000 non-critical data sets are determined from these 10,000 data sets, then when performing the aging operation, the 8,000 critical data sets are retained and the 2,000 non-critical data sets are deleted.

[0087] In an example of this embodiment, an aging operation can be performed on a time-series data segment multiple times. Refer to Figure 3 , it can be set that after an aging operation on the time-series data segment composed of 10,000 data sets acquired on October 22, the number of data sets in the time-series data segment is 8,000. Then, the time-series data segment is aged again. Specifically, the critical data sets and non-critical data sets in the time-series data segment can be determined by using a time-series data downsampling algorithm such as the PIP algorithm. It is set that 4,000 critical data sets and 4,000 non-critical data sets are determined from these 8,000 data sets. Then, when performing the aging operation, the 4,000 critical data sets are retained and the 4,000 non-critical data sets are deleted.

[0088] In an example of this embodiment, when performing an aging operation on a time series data segment again, the time series data segment and the time series data segments adjacent to the time series data segment can be merged into one time series data segment, and then, an aging operation is performed on the merged time series data segment. Refer to Figure 3 , the time series data segments with acquisition dates of October 16, October 17, October 18, October 19, October 20, and October 21 can be merged into a time series data segment containing 6,000 data sets. Then, using a time series data downsampling algorithm such as the PIP algorithm, the key data sets and non-key data sets in the merged time series data segment are determined. Assuming that 3,000 key data sets and 3,000 non-key data sets are determined from the merged time series data segment, then during the aging operation, the 3,000 key data sets are retained and the 3,000 non-key data sets are deleted.

[0089] In some embodiments, after performing an aging operation on a time series data segment, if the remaining data sets cannot reflect the data change trend, that is, the remaining data sets cannot be fitted, then the entire time series data segment is deleted. The fact that the data sets cannot reflect the data change trend indicates that the time series data segment is no longer valuable and can be completely deleted to free up the storage space of the storage space 200. For example, as Figure 3 shown, it can be set that after multiple aging operations, the time series data segment with an acquisition date of September has 1,500 remaining data sets. After performing an aging operation on the 1,500 data sets, the remaining data sets can no longer reflect the data change trend, and at this time, the entire time series data segment can be deleted.

[0090] The processing module 130 in the management device 100 can obtain the available capacity of the storage space 200, where the available capacity refers to the available capacity of the data sets used to store time series data in the storage space 200. Exemplarily, all the capacity of the storage space 200 is used to store the data sets of time series data, and the processing module 130 can count the used capacity of the storage space 200, and then, based on the total capacity and the used capacity of the storage space 200, calculate the available capacity.

[0091] In some embodiments, the processing module 130 can adjust the acquisition strategy for the processing module 110 to acquire the data sets of time series data based on the available capacity. Among them, the acquisition speed in the acquisition strategy is positively correlated with the available capacity.

[0092] In one example, the processing module 130 may continuously adjust the acquisition speed in the acquisition policy based on changes in the available capacity. In an example of this example, a maximum acquisition speed may be preset, and the ratio of the acquisition speed in the acquisition policy to the maximum acquisition speed is equal to the ratio of the available capacity to the total capacity. That is to say, the processing module 130 may calculate the ratio of the available capacity to the total capacity, and then multiply this ratio by the maximum acquisition speed to obtain the adjusted acquisition speed.

[0093] In one example, the processing module 130 may discretely adjust, i.e., stepwise adjust, the acquisition speed of the acquisition policy based on changes in the available capacity. Specifically, when the available capacity is lower than a preset threshold B1, the acquisition speed is decreased by a preset amplitude B11 (such as 10% of the current acquisition speed). When the available capacity is lower than a preset threshold B2, the acquisition speed is decreased again by a preset amplitude B21 (such as 20% of the current acquisition speed), and so on. Among them, the threshold B1 is greater than the threshold B2. In addition, when the available capacity is zero, the acquisition speed can be decreased to zero.

[0094] In some embodiments, the processing module 130 may adjust the aging policy for aging the data set of the time series data in the storage space 200 by the aging module 110 based on the available capacity. Among them, the aging speed in the aging policy is negatively correlated with the available capacity.

[0095] In one example, the processing module 130 may continuously adjust the aging speed in the aging policy based on changes in the available capacity. In an example of this example, a maximum aging speed may be preset, and the ratio of the aging speed in the aging policy to the maximum aging speed is negatively correlated with the ratio of the available capacity to the total capacity. Among them, the processing module 130 may calculate the ratio of the available capacity to the total capacity, then subtract this ratio from 1 to obtain the target ratio of the aging speed, and then multiply this target ratio by the maximum aging speed to obtain the adjusted aging speed.

[0096] In one example, the processing module 130 may discretely adjust the aging speed in the aging policy based on changes in the available capacity. Specifically, when the available capacity is lower than a preset threshold C1, the aging speed is increased by a preset amplitude C11 (such as 10% of the current aging speed). When the available capacity is lower than a preset threshold C2, the aging speed is increased again by a preset amplitude C21 (such as 20% of the current aging speed), and so on. Among them, the threshold C1 is greater than the threshold C2. In addition, when the available capacity is zero, the aging speed can be increased to the preset maximum aging speed.

[0097] In one example, the processing module 130 may determine whether a time-series data segment is an abnormal data segment. In one example, the similarity between the time-series data segment and known abnormal time-series data may be calculated. When the similarity is greater than a preset similarity threshold, it is confirmed that the time-series data segment is an abnormal time-series data segment. In one example, it may be determined whether the fluctuation amplitude of the time-series data segment is greater than a preset amplitude threshold. If it is greater, it is confirmed that the time-series data segment is an abnormal time-series data segment. In one example, a time-series data segment including an abnormal data set may be referred to as an abnormal time-series data segment. An abnormal data set refers to a data set including abnormal data. Among them, when the data in one or more data sets in the time-series data segment exceeds a preset range, it is confirmed that the data set is an abnormal data set, and further it is confirmed that the time-series data segment is abnormal time-series data.

[0098] In one example of this example, the abnormal time-series data segment has an importance level. Among them, the importance level of the abnormal time-series data segment is positively correlated with the similarity between the time-series data segment and known abnormal time-series data. Or, the importance level of the abnormal time-series data segment is positively correlated with the difference in fluctuation amplitude, where the difference in fluctuation amplitude is the difference between the fluctuation amplitude of the abnormal time-series data segment and the preset amplitude threshold. Or, the importance level of the abnormal time-series data segment is positively correlated with the proportion of abnormal data sets in the abnormal time-series data segment, and the proportion of abnormal data sets refers to the ratio of the number of abnormal data sets in the time-series data segment to the total number of data sets.

[0099] In some embodiments, the processing module 130 may feedback the identified abnormal time-series data segment to the acquisition module 110. The acquisition module 110 may, based on the abnormal time-series data segment, increase the acquisition priority level of the data source of the abnormal time-series data segment, and then, according to the acquisition speed corresponding to the increased acquisition priority level, acquire a data set of time-series data from the data source.

[0100] Exemplarily, the abnormal time-series data segment has an importance level. And different importance levels correspond to different improvement amplitudes, and the higher the importance level, the greater the improvement amplitude corresponding to the importance level. The acquisition module 110 may increase the acquisition priority level of the data source of the abnormal time-series data segment according to the improvement amplitude corresponding to the importance level of the abnormal time-series data segment. That is, the higher the importance level of the abnormal time-series data segment, the greater the improvement amplitude of the acquisition priority level of the data source of the abnormal time-series data segment.

[0101] In some embodiments, the processing module 130 may feedback the identified abnormal time-series data segment to the aging module 120. The aging module 120 may, based on the abnormal time-series data segment, reduce the aging priority level of the abnormal time-series data segment, and then, according to the aging speed corresponding to the reduced aging priority level, age the abnormal time-series period.

[0102] Exemplarily, the abnormal timing data segment has an importance level. Moreover, different importance levels correspond to different reduction amplitudes, and there is a positive correlation between the level of the importance level and the magnitude of the reduction amplitude corresponding to the importance level. The aging module 120 can reduce the aging priority level of the abnormal timing data segment according to the reduction amplitude corresponding to the importance level of the abnormal timing data segment. That is, the higher the importance level of the abnormal timing data segment, the greater the reduction amplitude of the aging priority level of the abnormal timing data segment.

[0103] In some embodiments, when the processing module 130 starts and initializes, it can detect the storage space 200. If the storage space 200 is not detected, the processing module 130 can create the storage space 200 in the storage pool connected to the management device 100. For example, it can select the disk with the optimal performance and create the storage space 200 based on the selected disk. The processing module 130 can map the disk of the storage space 200 to the management node 100. The processing module 130 can detect the health status of the storage space 200. If the storage space 200 is in a sub-healthy state, capacity overlimit, sudden failure state, or disk life expiration, etc., the processing module 130 can re-select the disk, create a new storage space 200 based on the re-selected disk, and map the new storage space 200 to the management node 100.

[0104] In some embodiments, referring to Figure 1 , the management device 100 further includes a display module 140. The display module 140 can display the timing data based on the data set of the timing data stored in the storage space 200. Among them, the user can, through a custom query method, select to query the timing data for a certain period of time and select what kind of timing data. In one example, referring to Figure 4 , the display module 140 can display a user interface through which the user can select the timing data for a certain period of time. For example, select the timing data for the period from 19:31 on June 1, 2021 to 22:31 on June 1, 2021. The user can, through this user interface, display what kind of timing data. For example, select the timing data corresponding to the number of input / output operations per second (IOPS) of the storage array.

[0105] In some embodiments, referring to Figure 1 , the management device 100 further includes an alarm module 150. The alarm module 150 can issue an alarm based on the abnormal timing data segment stored in the storage space 200.

[0106] In some embodiments, referring to Figure 1, the management device 100 further includes a download module 160. The user can download the time-series data from the storage space 200 through the download module 160.

[0107] The system architecture provided by the embodiments of the present application is introduced above. Next, in combination with this system architecture, the time-series data management method provided by the embodiments of the present application will be introduced.

[0108] This method can be executed by the management device 100. As Figure 5 shown, this method includes the following steps.

[0109] Step 501, the management device 100 obtains a data set of the time-series data and stores the obtained data set of the time-series data in the storage space 200.

[0110] After the management device 100 is started, it can obtain a data set of time-series data according to the initialized acquisition strategy. Among them, the initialized acquisition strategy can be implemented with reference to the above introduction and will not be elaborated here.

[0111] During the operation of the management device 100, based on the available capacity of the data set used by the management device 100 to store time-series data in the storage space 200, the acquisition speed in the acquisition strategy can be adjusted, that is, the acquisition speed of the management device 100 for obtaining the data set of time-limit data is adjusted. The management device 100 continues to obtain the data set of time-series data according to the adjusted acquisition speed. Among them, the acquisition speed is positively correlated with the available capacity of the storage space 200. In this way, when the available capacity of the storage space 200 is large, the data set of time-series data can be obtained at a relatively high speed. When the available capacity of the storage space 200 is small, the data set of time-series data can be obtained at a relatively low speed, thereby reducing the data stored in the storage space 200 and preventing the capacity of the storage space 200 from exceeding the limit.

[0112] In some embodiments, the management device 100 obtains the data set of the time-series data, including: collecting the data set for the target object to obtain the data set of the time-series data; wherein, the collection times of the data in different data sets of the time-series data are different.

[0113] Among them, the target object can be a module (software module or hardware module) in the computer system. That is to say, the management device 100 can collect data for one or more modules in the computer system to obtain a data set, and this data set can form time-series data. Through this time-series data, the health status of the module or the computer system can be evaluated, which is convenient for the management of the computer system.

[0114] In an example of this embodiment, the acquisition speed at which the management device 100 acquires the data set of the timing data includes the acquisition speed at which the management device 100 collects data from the target object. The management device 100 collecting data from the target object includes: the management device 100 collecting a data set from the target object at a first acquisition speed to obtain the data set of the timing data; when the data set of the timing data includes abnormal data, the management device 100 collecting a data set from the target object at a second acquisition speed; wherein, the second acquisition speed is greater than the first acquisition speed.

[0115] Collecting an abnormal data set indicates that the target object may have an abnormality. In this case, data is collected from the target object at a faster acquisition speed. Among them, a faster acquisition speed corresponds to a smaller acquisition granularity. By collecting data from the target object at a faster acquisition speed, fine-grained data of the target object can be obtained, facilitating a detailed analysis of the target object.

[0116] Exemplarily, it can be determined whether an abnormal data set is collected from the target object by determining whether the timing data segment collected from the target object is an abnormal timing data segment. The specific implementation can refer to the above introduction and will not be elaborated here.

[0117] In some embodiments, storing the obtained data set of the timing data in the storage space 200 includes: inputting the obtained data set of the timing data into a pre-trained encoding module, so that the encoding module outputs an encoding result of the obtained data set of the timing data; then, storing the encoding result in the storage space 200.

[0118] The encoding result is an encoding vector obtained by the encoding module extracting features from the data set of the timing data. The data volume of the encoding result is smaller than the data volume of the data set of the timing data, and the encoding result contains the features of the data set of the timing data. Storing the encoding result in the storage space 200 can not only store the features of the data set of the timing data, but also save the storage space of the storage space 200.

[0119] In addition, the more specific implementation of step 501 can refer to the above introduction of the acquisition module 110 and the processing module 130, and will not be elaborated one by one here.

[0120] Step 502, the management device 100 ages the data set of the timing data in the storage space 200.

[0121] After the management device 100 is started, the dataset of the timing data in the storage space 200 can be aged according to the initialized aging policy. The initialized aging policy can be implemented with reference to the above description and will not be elaborated here.

[0122] During the operation of the management device 100, based on the available capacity of the dataset in the storage space 200 for storing the timing data of the management device 100, the aging speed in the aging policy can be adjusted, and the aging speed of the dataset of the timing data in the storage space 200 aged by the management device 100 can be adjusted. The management device 100 ages the dataset of the timing data in the storage space 200 according to the adjusted aging speed. Among them, the aging speed is negatively correlated with the available capacity of the storage space 200. In this way, when the available capacity of the storage space 200 is large, the dataset of the timing data can be aged at a smaller speed, so that the dataset of the timing data can be stored for a long time. When the available capacity of the storage space 200 is small, the dataset of the timing data can be aged at a larger speed, so as to release the storage space occupied by the dataset as soon as possible and prevent the capacity of the storage space 200 from exceeding the limit.

[0123] In some embodiments, the management device 100 ages the dataset of the timing data in the storage space 200, including: the management device ages the dataset of the timing data in the storage space 200 at a first aging speed; when the dataset of the timing data in the storage space 200 includes an abnormal dataset, the management device ages the dataset of the timing data in the storage space 200 at a second aging speed; among them, the second aging speed is less than the first aging speed.

[0124] The dataset of the timing data including the abnormal dataset may specifically be that the timing data segment includes the abnormal dataset, or the timing data segment is an abnormal data segment. The timing data segment including the abnormal dataset or the abnormal data segment may reflect an abnormal situation and has great analysis value for analyzing the health status of the computer system. Aging the timing data segment including the abnormal dataset or the abnormal data segment at a small aging speed can make the timing data with great analysis value be stored for a longer time, which is convenient for analyzing the health status of the computer system.

[0125] In some embodiments, the timing data includes a key dataset and a non-key dataset. As described above, among them, the key dataset is the dataset in the timing data that reflects the change trend of the timing data, and the non-key dataset is the dataset in the timing data other than the key dataset. The management device 100 ages the dataset of the timing data in the storage space, including: aging the non-key dataset of the timing data in the storage space.

[0126] When performing an aging operation on time series data, the time series data segment can be used as the aging object. Specifically, Figure 6 As shown, the data sets in the time series data segment are divided into key data sets and non-key data sets by using time series data dynamic downsampling algorithms, such as the PIP algorithm, etc. Then, the non-key data sets are deleted from the storage space 200, and the key data sets are retained in the storage space 200. In this way, by deleting the non-key data sets, the storage space occupied by the time series data in the storage space 200 is reduced. By retaining the key data sets in the storage space 200, the time series data can still be used to evaluate the health status of the computer system, etc.

[0127] In addition, a more specific implementation of step 502 can refer to the above introduction to the aging module 120 and the processing module 130, which will not be described in detail here.

[0128] In some embodiments, the method includes: receiving a display indication for the time series data; in response to the display indication, supplementing the time series data with the non-critical data set; and displaying the time series data based on the supplemented non-critical data set and the data set of the time series data stored in the storage space 200 (i.e., the critical data set of the time series data).

[0129] As mentioned above, the aging operation of time series data is to delete non-critical data sets in the time series data. Figure 6 As shown, although the non-critical data set has little impact on the changing trend of the time series data, it can reflect the details of the time series data. When the time series data needs to be displayed, the data set can be supplemented for the time series data to supplement the non-critical data set deleted by the aging operation. Among them, the data set is supplemented for the time series data through data point supplementation algorithms such as upsampling algorithms and interpolation algorithms. Then, based on the supplemented data set (i.e., the non-critical data set) and the data set retained in the storage space 200 (i.e., the critical data set), the time series data is displayed, which can display time series data with richer details.

[0130] In some embodiments, the storage space 200 stores the encoding result of the data set of time series data. The encoding result is an encoding vector obtained by the encoding module performing feature extraction on the data set of time series data. The encoding module corresponds to a decoding module, and the decoding module is used to decode the encoding result into the data set of the time series data; the method includes: receiving a display instruction for the time series data; in response to the display instruction, inputting the encoding result to the decoding module so that the decoding module outputs the data set of the time series data; based on the data set of the time series data output by the decoding module, displaying the time series data.

[0131] Among them, the encoding module and the decoding module belong to the same AI model, and the encoding module and the decoding module can be obtained by training the AI ​​model. When the time series data needs to be displayed, the encoding result is decoded into a data set of the time series data through the decoding module, and then the time series data can be displayed based on the data set of the time series data. Thereby, the storage space of the storage space 200 is saved while ensuring that the time series data can be displayed normally.

[0132] To sum up, in the process of acquiring and storing time series data, the acquisition speed of the data set for acquiring the time series data and / or the aging speed of the data set of the time series data in the aging storage space can be dynamically adjusted based on the available capacity for storing the time series data, so that the acquisition speed and aging speed of the data set of the time series data match, so as to maintain the available capacity at a reasonable level, avoid capacity overruns, and ensure that there is capacity to store the latest data set of the time series data.

[0133] See also Figure 7 The present application embodiment also provides a time series data management device 700. Figure 7 As shown, the management device 700 includes:

[0134] An acquisition module 710 is used to acquire the data set of the time series data and store the acquired data set of the time series data in a storage space;

[0135] An aging module 720, configured to age the data set of the time series data in the storage space;

[0136] Among them, the acquisition speed of the data set of the time series data by the acquisition module 710 is positively correlated with the available capacity of the storage space, and / or the aging speed of the data set of the time series data in the storage space by the aging module 720 is negatively correlated with the available capacity; the available capacity is the available capacity in the storage space used by the management device to store the data set of the time series data.

[0137] In some embodiments, the acquisition speed is a collection speed; the acquisition module 710 is used to: collect data on the target object to obtain a data set of the time series data; wherein the collection time of data in different data sets of the time series data is different.

[0138] In an example of this embodiment, the acquisition speed includes the acquisition speed at which the acquisition module collects data on the target object; the acquisition module 710 is used to: collect data on the target object at a first acquisition speed; when the data set of the time series data includes abnormal data, collect data on the target object at a second acquisition speed; wherein the second acquisition speed is greater than the first acquisition speed.

[0139] In some embodiments, the aging module 720 is used to: age the data set of the time series data in the storage space at a first aging speed; when the data set of the time series data in the storage space includes abnormal data, age the data set of the time series data in the storage space at a second aging speed; wherein the second aging speed is less than the first aging speed.

[0140] In some embodiments, the time series data includes a critical data set and a non-critical data set; wherein, the critical data set is a data set in the time series data that reflects the changing trend of the time series data, and the non-critical data set is a data set in the time series data other than the critical data set; the aging module 720 is used to: age the non-critical data set of the time series data in the storage space.

[0141] In an example of this embodiment, the management device 700 further includes:

[0142] A first receiving module 730, configured to receive a display instruction for the time series data;

[0143] A supplementing module 740, configured to supplement the time series data with the non-critical data set in response to the display indication;

[0144] The first display module 750 is configured to display the time series data based on the supplemented non-critical data set and the data set of the time series data stored in the storage space.

[0145] In some embodiments, the acquisition module 710 is used to: input the data set of the time series data into a pre-trained encoding module so that the encoding module outputs the encoding result of the data set of the time series data; and store the encoding result into the storage space.

[0146] In an example of this embodiment, the encoding module corresponds to a decoding module, and the decoding module is used to decode the encoding result into a data set of the time series data; the management device 700 also includes:

[0147] A second receiving module 760, configured to receive a display instruction for the time series data;

[0148] An input module 770, configured to respond to the display instruction and input the encoding result to the decoding module, so that the decoding module outputs the data set of the time series data;

[0149] The second display module 780 is used to display the time series data based on the data set of the time series data output by the decoding module.

[0150] In addition, the functions of each functional module of the management device 700 can refer to the above description of Figure 1 , Figure 2 or Figure 5 The introduction and implementation of the illustrated embodiment will not be repeated here.

[0151] This embodiment of the application provides a computing device 800. Figure 8 As shown, the computing device 800 includes a processor 810 and a memory 820. The memory 820 is used to store files and executable programs. The processor 810 is used to execute the executable program stored in the memory 820, so that the computing device 800 can perform the operations performed by the management device 100 above, or Figure 5 The method shown.

[0152] The present application also provides a computer program product including instructions. The computer program product may be a software or program product including instructions that can be run on a computing device or stored in any available medium. When the computer program product is run on a computing device, the computing device is caused to execute Figure 5 The method shown.

[0153] The present application also provides a computer-readable storage medium. The computer-readable storage medium may be any available medium that can be stored by a computing device or a data storage device such as a data center that contains one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk). The computer-readable storage medium includes instructions that instruct the computing device to execute Figure 5 The method shown.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for managing time-series data, characterized in that The time-series data includes a data set generated in chronological order; the method includes: The management device acquires the data set of the time-series data and stores the acquired data set of the time-series data in a storage space; The management device ages the data set of the time-series data in the storage space; Wherein, the acquisition speed of the management device for acquiring the data set of the time-series data is positively correlated with the available capacity of the storage space, and / or, the aging speed of the management device for aging the data set of the time-series data in the storage space is negatively correlated with the available capacity; The available capacity is the available capacity in the storage space for the management device to store the data set of the time-series data.

2. The method according to claim 1, characterized in that The management device acquiring the data set of the time-series data includes: The management device performs data acquisition on a target object to obtain the data set of the time-series data; wherein, the acquisition times of the data in different data sets of the time-series data are different.

3. The method according to claim 2, wherein The acquisition speed includes the acquisition speed of the management device for performing data acquisition on the target object; The management device performing data acquisition on a target object includes: The management device performs data acquisition on the target object at a first acquisition speed; When the data set of the time-series data includes abnormal data, the management device performs data acquisition on the target object at a second acquisition speed; wherein, the second acquisition speed is greater than the first acquisition speed.

4. The method according to any one of claims 1 to 3, characterized in that The management device aging the data set of the time-series data in the storage space includes: The management device ages the data set of the time-series data in the storage space at a first aging speed; When the data set of the time-series data in the storage space includes abnormal data, the management device ages the data set of the time-series data in the storage space at a second aging speed; Wherein, the second aging speed is less than the first aging speed.

5. The method according to any one of claims 1-4, characterized in that, The time-series data includes a key data set and a non-key data set; wherein, the key data set is the data set in the time-series data that reflects the change trend of the time-series data, and the non-key data set is the data set in the time-series data other than the key data set; the management device aging the data set of the time-series data in the storage space includes: aging the non-key data set of the time-series data in the storage space.

6. The method according to claim 5, characterized in that, The method includes: Receiving a display instruction for the time-series data; In response to the display instruction, supplementing the non-key data set for the time-series data; Based on the supplemented non-key data set and the data set of the time-series data stored in the storage space, displaying the time-series data.

7. The method according to any one of claims 1-4, characterized in that, The storing the acquired data set of the time-series data in the storage space includes: Inputting the data set of the time-series data into a pre-trained encoding module, so that the encoding module outputs an encoding result of the data set of the time-series data; Storing the encoding result in the storage space.

8. The method according to claim 7, wherein The encoding module corresponds to a decoding module, and the decoding module is used to decode the encoding result into a data set of the timing data; the method includes: Receiving a display instruction for the timing data; In response to the display instruction, inputting the encoding result into the decoding module so that the decoding module outputs a data set of the timing data; Based on the data set of the timing data output by the decoding module, displaying the timing data.

9. A time series data management device, characterized in that, The timing data includes data sets generated in chronological order; the management device includes: An acquisition module, configured to acquire a data set of the timing data and store the acquired data set of the timing data in a storage space; An aging module, configured to age the data set of the timing data in the storage space; Wherein, the acquisition speed of the acquisition module for acquiring the data set of the timing data is positively correlated with the available capacity of the storage space, and / or, the aging speed of the aging module for aging the data set of the timing data in the storage space is negatively correlated with the available capacity; The available capacity is the available capacity in the storage space for the management device to store the data set of the timing data.

10. The management device according to claim 9, characterized in that, The acquisition module is used for: performing data acquisition on a target object to obtain a data set of the timing data; wherein, the acquisition times of the data in different data sets of the timing data are different.

11. The management device according to claim 10, wherein The acquisition speed includes the acquisition speed of the acquisition module for performing data acquisition on the target object; the acquisition module is used for: Performing data acquisition on the target object at a first acquisition speed; When the data set of the timing data includes abnormal data, performing data acquisition on the target object at a second acquisition speed; wherein, the second acquisition speed is greater than the first acquisition speed.

12. The management device according to any one of claims 9-11, characterized in that, The aging module is used for: Aging the data set of the timing data in the storage space at a first aging speed; When the data set of the timing data in the storage space includes abnormal data, aging the data set of the timing data in the storage space at a second aging speed; Wherein, the second aging speed is less than the first aging speed.

13. The management device according to any one of claims 9-12, characterized in that, The timing data includes a key data set and a non-key data set; wherein, the key data set is a data set in the timing data that reflects the change trend of the timing data, and the non-key data set is a data set in the timing data other than the key data set; the aging module is used for: aging the non-key data set of the timing data in the storage space.

14. The management device according to claim 13, wherein The management device further includes: A first receiving module, configured to receive a display instruction for the timing data; A supplement module, configured to, in response to the display instruction, supplement the non-key data set for the timing data; A first display module, configured to display the timing data based on the supplemented non-key data set and the data set of the timing data stored in the storage space.

15. The management device according to any one of claims 9-12, characterized in that, The acquisition module is used for: Inputting the data set of the timing data into a pre-trained encoding module so that the encoding module outputs an encoding result of the data set of the timing data; Store the encoding result in the storage space.

16. The management device according to claim 15, wherein The encoding module corresponds to a decoding module, and the decoding module is used to decode the encoding result into a data set of the timing data; the management device further includes: A second receiving module, configured to receive a display instruction for the timing data; An input module, configured to, in response to the display instruction, input the encoding result into the decoding module, so that the decoding module outputs the data set of the timing data; A second display module, configured to display the timing data based on the data set of the timing data output by the decoding module.

17. A computing device, characterized in that, Comprising: A memory for storing an executable program; A processor for executing the method according to any one of claims 1-8 by running the executable program.

18. A computer-readable storage medium, characterized in that, Comprising computer program instructions, when the computer program instructions are executed by a computing device, the computing device executes the method according to any one of claims 1-8.

19. A computer program product comprising instructions, characterized in that, When the instructions are run by a computer device, the computer device is caused to execute the method according to any one of claims 1-8.

Images