CN116991448B - Operation and maintenance time window detection method and system of server, storage medium and server - Google Patents
Operation and maintenance time window detection method and system of server, storage medium and server Download PDFInfo
- Publication number
- CN116991448B CN116991448B CN202311245830.6A CN202311245830A CN116991448B CN 116991448 B CN116991448 B CN 116991448B CN 202311245830 A CN202311245830 A CN 202311245830A CN 116991448 B CN116991448 B CN 116991448B
- Authority
- CN
- China
- Prior art keywords
- index data
- period
- power consumption
- server
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 68
- 238000001514 detection method Methods 0.000 title claims abstract description 44
- 230000000737 periodic effect Effects 0.000 claims abstract description 109
- 238000000034 method Methods 0.000 claims abstract description 43
- 238000004422 calculation algorithm Methods 0.000 claims description 102
- 230000008859 change Effects 0.000 claims description 67
- 238000012795 verification Methods 0.000 claims description 52
- 238000004364 calculation method Methods 0.000 claims description 51
- 238000012545 processing Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 23
- 238000010606 normalization Methods 0.000 claims description 16
- 230000015654 memory Effects 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 9
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/60—Software deployment
- G06F8/65—Updates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9035—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/70—Software maintenance or management
- G06F8/71—Version control; Configuration management
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention provides a method, a system, a storage medium and a server for detecting an operation and maintenance time window of a server, which relate to the field of servers and comprise the following steps: collecting index data of a server; determining whether the server is in an idle state according to the periodic variation state of the index data; if the server is not in the idle state, performing periodic analysis on the index data to obtain a periodic analysis result; and determining the operation and maintenance time window according to the period analysis result. After the index data of the server are collected, the invention sequentially realizes the no-load analysis and the period analysis of the server through the analysis and the detection of the index data, performs the analysis layer by layer, finally analyzes the optimal operation and maintenance time window, can accurately determine the operation and maintenance time window for the server with strong period, weak period and even no period, reduces the interruption of the service and ensures the reliable operation of the service.
Description
Technical Field
The present invention relates to the field of servers, and in particular, to a method and a system for detecting an operation and maintenance time window of a server, a storage medium, and a server.
Background
In the operation and maintenance of a data center, security holes such as an operating system, software, middleware and the like or new versions are frequently released and need to be upgraded, however, short-time interruption of the system or reduction of service processing speed is caused by more or less system upgrading, and for the selection of operation and maintenance time windows for operations such as service upgrading, middleware upgrading, operating system patching and the like, the upgrading is often performed according to the experience of a service manager or a fixed time period constrained by a company, and the time windows are selected in an empirical mode, so that the peak period of the service can be avoided, the upgrading and maintenance can be performed in the low-peak period, and the time windows with the minimum service influence are not the time windows.
At present, statistical analysis can be performed in an hour statistical mode, so that an operation and maintenance time window is recommended, and automatic service upgrading is performed. If the business is not periodically changed according to the day, the statistical analysis is inaccurate, and the business can be influenced.
Therefore, how to accurately detect the operation and maintenance time window of the server is a technical problem that needs to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide a method and a system for detecting an operation and maintenance time window of a server, a computer readable storage medium and the server, which can accurately detect the operation and maintenance time window of the server.
In order to solve the technical problems, the invention provides an operation and maintenance time window detection method of a server, which comprises the following specific technical scheme:
collecting index data of a server;
determining whether the server is in an idle state according to the periodic variation state of the index data;
if the server is not in the idle state, performing periodic analysis on the index data to obtain a periodic analysis result;
and determining the operation and maintenance time window according to the period analysis result.
In one aspect, the index data of the acquisition server includes:
and collecting the overall power consumption of the server and/or the utilization rate of a system processor.
In one aspect, collecting the overall power consumption and/or the system processor utilization of the server includes:
and collecting the overall power consumption of the server in an out-of-band manner and/or collecting the utilization rate of the system processor in an in-band manner.
In one aspect, the collecting the overall power consumption and/or the system processor utilization of the server includes:
and acquiring the overall power consumption of the server and/or the utilization rate of a system processor according to the set minimum acquisition frequency.
In one aspect, when the overall power consumption of the server and/or the utilization rate of the system processor are collected, the method further comprises:
If the utilization rate of the system processor is acquired, stopping acquiring the overall power consumption of the server;
and if the utilization rate of the system processor is not acquired, acquiring the overall power consumption of the server, and carrying out normalization processing on the overall power consumption.
In one aspect, normalizing the overall power consumption includes:
reading the power consumption of the whole machine; the power consumption of the whole machine comprises all acquired data in a set time length;
determining a maximum power consumption value and a minimum power consumption value in the power consumption of the whole machine;
and determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value.
In one aspect, the determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value includes:
taking a larger value between the maximum power consumption value and the minimum power consumption value which is preset times as a large power consumption value, and taking the minimum power consumption value as a small power consumption value;
and taking the ratio of the difference value between the current power consumption and the small power consumption value to the difference value between the large power consumption value and the small power consumption value as normalized power consumption data.
In one aspect, the determining whether the server is in an idle state according to the periodically changing state of the index data includes:
Calculating the average value of the index data;
substituting the average value into an idle load calculation formula to obtain an idle load calculation result;
if the idle load calculation result is not smaller than a preset value, confirming that the server is in an idle load state;
and if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state.
In one aspect, the no-load calculation formula is:
score=(0.1-Avg)*A+ (0.15-Avg99)*（1-A）；
wherein score is an idle calculation result, avg99 represents the average value of the index data of the first 99% after the index data is sequenced, and a is a coefficient.
In one aspect, performing a period analysis on the index data to obtain a period analysis result includes:
determining a periodic variation of the index data;
if the index data is changed in a strong period, counting the index data according to the strong period to obtain a period analysis result;
if the index data is in weak periodic variation, predicting the index data by adopting a periodic prediction algorithm to obtain a periodic analysis result;
and if the index data does not change periodically, predicting the index data by adopting a non-periodicity prediction algorithm to obtain a period analysis result.
In one aspect, the determining the periodic variation of the index data includes:
Performing autocorrelation calculation on the index data, and determining a periodic variation index;
if the intra-period change index is larger than a first threshold value, confirming that the index data is in strong period change;
if the intra-period change index is not greater than the first threshold and is greater than the second threshold, confirming that the index data is in weak period change;
and if the periodic variation index is not greater than the second threshold value, confirming that the index data does not have periodic variation.
In one aspect, the performing autocorrelation calculation on the index data, and determining the intra-period variation index includes:
calculating the number of the acquisition points in a preset time length;
squaring and summing index data acquired by all the points in the preset time length to obtain a first correlation value;
multiplying the index data acquired in the preset time period by the index data acquired in the next preset time period in a one-to-one correspondence manner to obtain a second correlation value;
calculating a correlation from the first correlation value and the second correlation value;
and determining a period change index according to the maximum amplitude, the period change amplitude and the correlation.
In one aspect, before determining the intra-period variation index according to the maximum amplitude, the period variation amplitude and the correlation, the method further includes:
And performing fast Fourier transform on the index data to obtain the maximum amplitude and the periodic variation amplitude.
In one aspect, the counting the index data according to the strong period to obtain a period analysis result includes:
and taking a time period corresponding to the time period when the average value of the index data in the strong period is minimum as a recommended time period of the operation and maintenance time window.
In one aspect, the predicting the index data by using a periodic prediction algorithm, and obtaining the periodic analysis result includes:
dividing the index data according to a set proportion to obtain training data and verification data;
predicting the training data by adopting a plurality of period prediction algorithms;
performing accuracy verification on each period prediction algorithm by using the verification data, and determining an optimal period prediction algorithm according to the symmetric average absolute percentage error;
and predicting the index data by using the optimal period prediction algorithm to obtain a period analysis result.
In one aspect, the predicting the index data by using the non-periodicity prediction algorithm, and obtaining the period analysis result includes:
dividing the index data according to a set proportion to obtain training data and verification data;
Predicting the training data by adopting a plurality of non-periodic prediction algorithms;
performing accuracy verification on each of the non-periodic prediction algorithms by using the verification data, and determining an optimal non-periodic prediction algorithm according to the symmetric average absolute percentage error;
and predicting the index data by using the optimal non-periodic prediction algorithm to obtain a periodic analysis result.
In one aspect, the performing accuracy verification on each of the periodic prediction algorithms using the verification data, determining an optimal periodic prediction algorithm based on a symmetric average absolute percentage error includes:
calculating a symmetrical average absolute percentage error of each period prediction algorithm for carrying out accuracy verification on the verification data, and taking a corresponding target period prediction algorithm as an optimal period prediction algorithm if the minimum value of the symmetrical average absolute percentage error is smaller than an error preset value;
in one aspect, if the minimum value of the symmetric average absolute percentage error is not less than the error preset value, the method further includes:
and counting the index data according to the strong period to obtain a period analysis result.
In one aspect, the calculating the mean value of the index data includes:
and sequencing the index data, and taking the index data before the fixed proportion position to calculate the average value.
In one aspect, sorting the index data, and calculating the average value of the index data before the fixed proportion position comprises:
and determining a reliable data coefficient, and determining the fixed proportion according to the reliable data coefficient.
The invention also provides an operation and maintenance time window detection system of the server, which comprises:
the data acquisition module is used for acquiring index data of the server;
the no-load detection module is used for determining whether the server is in an no-load state according to the periodic variation state of the index data;
the period analysis module is used for carrying out period analysis on the index data if the server is not in the idle state, so as to obtain a period analysis result;
and the time window detection module is used for determining the operation and maintenance time window according to the period analysis result.
In one aspect, the data acquisition module is a module for acquiring overall power consumption of the server and/or utilization rate of a system processor.
In one aspect, the data acquisition module is a module for out-of-band acquisition of the overall power consumption of the server and/or in-band acquisition of the utilization rate of the system processor.
In one aspect, the data acquisition module is a module for acquiring the overall power consumption of the server and/or the utilization rate of a system processor according to a set minimum acquisition frequency.
In one aspect, the data acquisition module further comprises:
the acquisition detection unit is used for stopping acquiring the overall power consumption of the server if the utilization rate of the system processor is acquired; and if the utilization rate of the system processor is not acquired, acquiring the overall power consumption of the server, and carrying out normalization processing on the overall power consumption.
In one aspect, an acquisition detection unit includes:
the normalization unit is used for reading the power consumption of the whole machine; the power consumption of the whole machine comprises all acquired data in a set time length; determining a maximum power consumption value and a minimum power consumption value in the power consumption of the whole machine; and determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value.
In one aspect, the normalization processing unit includes:
the normalization processing unit is used for taking a larger value between the maximum power consumption value and the minimum power consumption value which is preset times as a power consumption large value and taking the minimum power consumption value as a power consumption small value; and taking the ratio of the difference value between the current power consumption and the small power consumption value to the difference value between the large power consumption value and the small power consumption value as normalized power consumption data.
In one aspect, the no-load detection module comprises:
the average value calculation unit is used for calculating the average value of the index data;
The no-load calculation unit is used for substituting the average value into a no-load calculation formula to obtain a no-load calculation result;
the no-load detection unit is used for confirming that the server is in a no-load state if the no-load calculation result is not smaller than a preset value; and if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state.
In one aspect, the period analysis module includes:
a period change analysis unit configured to determine a period change of the index data; if the index data is changed in a strong period, counting the index data according to the strong period to obtain a period analysis result; if the index data is in weak periodic variation, predicting the index data by adopting a periodic prediction algorithm to obtain a periodic analysis result; and if the index data does not change periodically, predicting the index data by adopting a non-periodicity prediction algorithm to obtain a period analysis result.
In one aspect, the period change analysis unit includes:
the periodic variation detection unit is used for carrying out autocorrelation calculation on the index data and determining a periodic variation index; if the intra-period change index is larger than a first threshold value, confirming that the index data is in strong period change; if the intra-period change index is not greater than the first threshold and is greater than the second threshold, confirming that the index data is in weak period change; and if the periodic variation index is not greater than the second threshold value, confirming that the index data does not have periodic variation.
In one aspect, the periodic variation detection unit includes:
the autocorrelation calculating unit is used for calculating the number of the acquisition points in the preset duration; squaring and summing index data acquired by all the points in the preset time length to obtain a first correlation value; multiplying the index data acquired in the preset time period by the index data acquired in the next preset time period in a one-to-one correspondence manner to obtain a second correlation value; calculating a correlation from the first correlation value and the second correlation value; and determining a period change index according to the maximum amplitude, the period change amplitude and the correlation.
In one aspect, the period change detection unit further includes:
and the data processing unit is used for carrying out fast Fourier transform on the index data to obtain the maximum amplitude and the periodical change amplitude.
In one aspect, the period change analysis unit includes:
and the strong period analysis unit is used for taking a time period corresponding to the time period when the average value of the index data in the strong period is minimum as a recommended time period of the operation and maintenance time window.
In one aspect, the period change analysis unit includes:
the weak period analysis unit is used for dividing the index data according to a set proportion to obtain training data and verification data; predicting the training data by adopting a plurality of period prediction algorithms; performing accuracy verification on each period prediction algorithm by using the verification data, and determining an optimal period prediction algorithm according to the symmetric average absolute percentage error; and predicting the index data by using the optimal period prediction algorithm to obtain a period analysis result.
In one aspect, the period change analysis unit includes:
the non-period analysis unit is used for dividing the index data according to a set proportion to obtain training data and verification data; predicting the training data by adopting a plurality of non-periodic prediction algorithms; performing accuracy verification on each of the non-periodic prediction algorithms by using the verification data, and determining an optimal non-periodic prediction algorithm according to the symmetric average absolute percentage error; and predicting the index data by using the optimal non-periodic prediction algorithm to obtain a periodic analysis result.
On the one hand, the weak period analysis unit is a unit for calculating a symmetric average absolute percentage error of each period prediction algorithm for verifying the accuracy of the verification data, and if the minimum value of the symmetric average absolute percentage error is smaller than the error preset value, the corresponding target period prediction algorithm is used as an optimal period prediction algorithm;
in one aspect, the mean value calculating unit is a unit for sorting the index data and calculating a mean value of the index data before the fixed proportion position is taken.
In one aspect, the method further comprises:
and the proportion determining unit is used for determining a reliable data coefficient and determining the fixed proportion according to the reliable data coefficient.
The invention also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor implements the steps of the method as described above.
The invention also provides a server comprising a memory in which a computer program is stored and a processor which when calling the computer program in the memory implements the steps of the method as described above.
The invention provides a method for detecting an operation and maintenance time window of a server, which comprises the following steps: collecting index data of a server; determining whether the server is in an idle state according to the periodic variation state of the index data; if the server is not in the idle state, performing periodic analysis on the index data to obtain a periodic analysis result; and determining the operation and maintenance time window according to the period analysis result.
After the index data of the server are collected, the invention sequentially realizes the no-load analysis and the period analysis of the server through the analysis and the detection of the index data, performs the analysis layer by layer, finally analyzes the optimal operation and maintenance time window, can accurately determine the operation and maintenance time window for the server with strong period, weak period and even no period, reduces the interruption of the service and ensures the reliable operation of the service.
The invention also provides an operation and maintenance time window detection system of the server, a computer readable storage medium and the server, which have the beneficial effects and are not repeated here.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for detecting an operation and maintenance time window of a server according to an embodiment of the present invention;
FIG. 2 is a flowchart of performing autocorrelation calculation on index data according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an operation and maintenance time window detection system of a server according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a server according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, fig. 1 is a flowchart of a method for detecting an operation and maintenance time window of a server according to an embodiment of the present invention, where the method includes:
s101: collecting index data of a server;
s102: determining whether the server is in an idle state according to the periodic variation state of the index data;
s103: if the server is not in the idle state, performing periodic analysis on the index data to obtain a periodic analysis result;
s104: and determining the operation and maintenance time window according to the period analysis result.
The index data of how to collect the service is not limited here, and the specific content of the index data is not limited. The index data may be data for feeding back the running state of the server, including but not limited to the overall power consumption of the server and/or the utilization rate of a system processor.
If the system is collected simultaneously, the power consumption of the whole server can be collected out of band, and/or the utilization rate of the system processor can be collected in band. The out-of-band acquisition can utilize the baseboard management controller of the server to acquire the overall power consumption of the server, and the in-band acquisition system processor utilization rate can directly utilize the service system instruction of the server to acquire the real-time system processor utilization rate. In addition, the specific collection frequency of the index data is not limited in this embodiment, and the overall power consumption of the server and/or the utilization rate of the system processor can be collected according to the set lowest collection frequency. The set minimum acquisition frequency may be set by the person skilled in the art, and is not limited herein.
In a possible implementation manner, when the overall power consumption of the server and/or the utilization rate of the system processor are collected, if the utilization rate of the system processor can be collected, the collection of the overall power consumption of the server can be stopped. And if the utilization rate of the system processor is not acquired, acquiring the overall power consumption of the server, and carrying out normalization processing on the overall power consumption. The utilization rate of the system processor is preferentially used, if the utilization rate of the system processor cannot be acquired, the utilization rate of the system processor can be considered to be further used for processing, and the server power consumption can be used after normalization processing is needed.
Here, there is no limitation on how the normalization process is performed, and the following is a possible scheme provided in this embodiment:
the first step, the power consumption of the whole machine is read; the power consumption of the whole machine comprises all acquired data in a set time length;
step two, determining a maximum power consumption value and a minimum power consumption value in the power consumption of the whole machine;
and thirdly, determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value.
When the third step is specifically performed, a larger value between the maximum power consumption value and the minimum power consumption value of a preset multiple may be taken as a large power consumption value, and the minimum power consumption value may be taken as a small power consumption value. And taking the ratio of the difference value between the current power consumption and the small power consumption value to the difference value between the large power consumption value and the small power consumption value as normalized power consumption data. The preset times are not limited and may be set by one skilled in the art.
For a clearer description of the normalization process, the following description of the above process is made using the corresponding formulas:
setting a maximum power consumption value ObMax and a minimum power consumption value ObMin of the acquired server power consumption, and executing the following calculation:
PowerMax=ObMax>2*ObMin? ObMax:2*ObMin。
PowerMin=ObMin。
metric_nor=(power-PowerMin)/( PowerMax- PowerMin)。
the ObMax >2×obmin: "previous result, namely maximum power consumption value ObMax as power consumption large value PowerMax, if not, execute": the "post result, namely 2 times the minimum power consumption value ObMin as the power consumption large value PowerMax. And the power consumption small value PowerMin is the minimum power consumption value ObMin, and the metric_nor refers to the normalized power consumption data.
And taking the ratio of the difference value between the current power consumption and the small power consumption value and the difference value between the large power consumption value and the small power consumption value as normalized power consumption data, wherein power refers to current index data.
After the index data are collected, whether the server is in an idle state or not is determined according to the periodic change state of the index data. Specifically, the average value of the index data can be calculated, and the average value is substituted into the no-load calculation formula to obtain a no-load calculation result. And if the idle load calculation result is not smaller than the preset value, confirming that the server is in the idle load state, and if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state.
The no-load calculation formula is not particularly limited, and in one possible embodiment, the following no-load calculation formula may be used:
score=(0.1-Avg)*A+ (0.15-Avg99)*（1-A）；
wherein score is an idle calculation result, avg99 represents the average value of the index data of the first 99% after the index data is sequenced, and a is a coefficient, which can be set by a person skilled in the art.
If A is 0.5, the corresponding no-load calculation formula is:
score=(0.1-Avg)*0.5+ (0.15-Avg99)*0.5。
when the index data is subjected to periodic analysis to obtain a periodic analysis result, the periodic variation of the index data can be determined first, and a corresponding analysis mode is executed according to the periodic variation:
if the index data is changed in a strong period, counting the index data according to the strong period to obtain a period analysis result;
if the index data is in weak periodic variation, predicting the index data by adopting a periodic prediction algorithm to obtain a periodic analysis result;
and if the index data does not change periodically, predicting the index data by adopting a non-periodicity prediction algorithm to obtain a period analysis result.
The method is not limited to how to determine the periodic variation of the index data, and the periodic variation index can be determined according to the autocorrelation calculation of the index data, and then the corresponding periodic variation can be determined according to the periodic variation index:
If the intra-period change index is larger than a first threshold value, confirming that the index data is in strong period change;
if the intra-period change index is not greater than the first threshold and is greater than the second threshold, confirming that the index data is in weak period change;
and if the periodic variation index is not greater than the second threshold value, confirming that the index data does not have periodic variation.
Referring to fig. 2, fig. 2 is a flowchart of an autocorrelation calculation for index data according to an embodiment of the present invention, and a process for autocorrelation calculation for index data according to the present embodiment may be as follows:
s201: calculating the number of the acquisition points in a preset time length;
s202: squaring and summing index data acquired by all the points in the preset time length to obtain a first correlation value;
s203: multiplying the index data acquired in the preset time period by the index data acquired in the next preset time period in a one-to-one correspondence manner to obtain a second correlation value;
s204: calculating a correlation from the first correlation value and the second correlation value;
s205: and determining a period change index according to the maximum amplitude, the period change amplitude and the correlation.
In a possible embodiment, before performing step S205 to determine the intra-period change index according to the maximum amplitude, the period change amplitude, and the correlation, a fast fourier transform may be further performed on the index data to obtain the maximum amplitude and the period change amplitude.
The above procedure is described below by way of example, assuming that the preset duration is 24 hours, and the number of collection points in 24 hours is determined, then the number of points length=24×60 (here assumed to be collected once a minute). And then squaring and summing the collected index data from 1 to length to obtain a first correlation value P1.
And multiplying the acquired index data 1 to length by the index data of the corresponding length+1 to 2 x length in a one-to-one correspondence manner to obtain a second correlation value P2.
The correlation corr=p2/P1 can be calculated from the first correlation value P1 and the second correlation value P2.
Finally, calculating 24-hour time-varying index: score=am24/amax+corr2-2.
Where Am24 is the magnitude of the 24 hour change and amax is the maximum magnitude. If score >0, then consider a 24 hour strong cycle change and use a strong cycle change prediction; if score > -0.1, then consider a weak periodic variation, then predict with a weak periodic variation; if score < = -0.1, no period change prediction is employed.
If the index data are counted according to the strong period, a time period corresponding to the time period when the average value of the index data in the strong period is minimum can be used as a recommended time period of the operation and maintenance time window.
If the index data is predicted by adopting a periodic prediction algorithm, the index data can be segmented according to a set proportion to obtain training data and verification data, then the training data is predicted by adopting a plurality of periodic prediction algorithms, accuracy verification is carried out on each periodic prediction algorithm by utilizing the verification data, an optimal periodic prediction algorithm is determined according to a symmetrical average absolute percentage error, and finally the index data is predicted by utilizing the optimal periodic prediction algorithm to obtain a periodic analysis result.
Similarly, when the index data is predicted by adopting a non-periodic prediction algorithm, the index data can be segmented according to a set proportion to obtain training data and verification data, the training data is predicted by adopting a plurality of non-periodic prediction algorithms, accuracy verification is performed on each non-periodic prediction algorithm by utilizing the verification data, an optimal non-periodic prediction algorithm is determined according to a symmetrical average absolute percentage error, and finally the index data is predicted by utilizing the optimal non-periodic prediction algorithm to obtain a periodic analysis result. The setting ratio is not limited, and may be set by those skilled in the art, and generally the training data amount is larger than the verification data amount.
If the verification data is used for executing accuracy verification on each period prediction algorithm, when an optimal period prediction algorithm is determined according to the symmetric average absolute percentage error, the symmetric average absolute percentage error of each period prediction algorithm for executing accuracy verification on the verification data can be calculated, and if the minimum value of the symmetric average absolute percentage error is smaller than an error preset value, the corresponding target period prediction algorithm is used as the optimal period prediction algorithm;
In the weak periodic pattern, the data is considered to have a certain periodic trend, but not a strong periodic like 24-hour periodic, where predictions are made using a predictive algorithm with a period, for example:
dividing the index data into 7:3,7 parts as training data and 3 parts as verification data.
With 7 pieces of data, predictions are made using various cycle prediction algorithms, such as propset (a time prediction algorithm), lstm (Long Short Term Memory networks, long and short term memory network) and the like. 3 data are used for accuracy verification, symmetrical average absolute percentage error smape can be used for algorithm accuracy verification, symmetrical average absolute percentage error smape of different period prediction algorithms is ordered, a minimum value S_min of the smape is obtained, if S_min is less than 0.1, the prediction accuracy of the algorithm for the S_min is considered to meet the requirement, and prediction analysis can be performed by using the algorithm. At this time, the algorithm is adopted, all index data are imported into the algorithm model, the model is retrained for prediction, and the time point of the predicted minimum value can be regarded as the optimal operation and maintenance time window. And if S_min > =0.1, counting the index data according to the strong period to obtain a period analysis result.
Similarly, if no periodic variation is determined, the index data may be segmented, for example, the index data may be similarly divided into 7:3,7 parts as training data and 3 parts as verification data. With 7 data, predictions are made using a non-periodic prediction algorithm, such as three-time smoothing, arima (Autoregressive Integrated Moving Average Model, differential autoregressive moving average model), and the like.
3 parts of data are used for accuracy verification, symmetrical average absolute percentage error smape is used for algorithm accuracy verification, smapes of different algorithms are ordered, a minimum value S_min of the smape is obtained, if S_min is less than 0.1, the prediction accuracy of the algorithm for the S_min is considered to meet the requirement, and prediction analysis can be performed by using the algorithm. At this time, the algorithm is adopted, all index data are imported into the algorithm model, the model is retrained for prediction, and the time point of the predicted minimum value can be regarded as the optimal operation and maintenance time window.
According to the embodiment of the invention, after the index data of the server are acquired, the no-load analysis and the period analysis of the server are sequentially realized through the analysis and the detection of the index data, the analysis is carried out layer by layer, the optimal operation and maintenance time window is finally analyzed, the operation and maintenance time window can be accurately determined for the server with strong period, weak period and even no period, the interruption of the service is reduced, and the reliable operation of the service is ensured.
On the basis of the embodiment, the index data may be ranked when the average value of the index data is calculated, and the average value is calculated by taking the index data before the fixed proportion position.
For example, after the index data is collected, the index data at all times may be sorted in order from small to large, the value of the position of 0.99 of the total length of the data is taken to obtain Avg99, and then the index data smaller than Avg99 is averaged to obtain Avg. In the present embodiment, the fixed ratio is not limited, and the fixed ratio may be set to a large value if the reliable data coefficient for the index data is considered to be reliable, but is generally smaller than 1.
It will be readily appreciated that the reliable data coefficients may also be determined before sorting the index data, taking the index data before a fixed scale position, and determining the fixed scale from the reliable data coefficients.
It can be seen that the embodiment of the invention mainly comprises four processes of index acquisition and processing, no-load analysis, period analysis and operation and maintenance time window analysis. The index collection and processing mainly collects data of the in-band processor utilization rate and the out-of-band server power consumption, the priority use of the processor utilization rate can be set, if the processor utilization rate cannot be collected, the server power consumption is used for processing, and the server power consumption can be used after normalization is needed. And carrying out no-load analysis treatment after normalization. And the no-load analysis is used for analyzing whether the server is in no-load or not by adopting a statistical analysis mode for the processed indexes, if the server is in no-load, the server can be upgraded at any moment, and otherwise, the server enters the period analysis. The period analysis means that the period of the data is analyzed, whether the period is changed in a strong period of 24 hours or not, if the period is changed in a strong period, the period is considered to be changed in a strong period, statistics is carried out according to the strong period transformation, and an operation and maintenance time window is analyzed; if the periodicity is very weak, for example, not 24 hours, a plurality of algorithms with periodicity prediction are adopted for prediction, whether the accuracy of the best algorithm meets the requirement is judged, if yes, the algorithm is adopted for predicting a future operation and maintenance time window, and if not, the operation and maintenance time window is statistically analyzed; if the period change does not exist, a plurality of asexual prediction algorithms are adopted for prediction, whether the accuracy of the best algorithm meets the requirement is judged, if yes, the algorithm is adopted for predicting a future operation and maintenance time window, and if not, the operation and maintenance time window is statistically analyzed. By stepwise analysis, the optimal operation and maintenance time window is determined, and the influence of service interruption caused by operation and maintenance time can be reduced, so that the continuity of service is ensured.
The following describes an operation and maintenance time window detection system of a server according to an embodiment of the present invention, where the operation and maintenance time window detection system described below and the operation and maintenance time window detection method of the server described above may be referred to correspondingly.
Referring to fig. 3, fig. 3 is a schematic structural diagram of an operation and maintenance time window detection system of a server according to an embodiment of the present invention, and the invention further provides an operation and maintenance time window detection system of a server, including:
the data acquisition module is used for acquiring index data of the server;
the no-load detection module is used for determining whether the server is in an no-load state according to the periodic variation state of the index data;
the period analysis module is used for carrying out period analysis on the index data if the server is not in the idle state, so as to obtain a period analysis result; the period analysis is used for determining the period change of the index data and predicting by adopting a corresponding prediction algorithm based on the period change;
and the time window detection module is used for determining the operation and maintenance time window according to the period analysis result.
Based on the foregoing embodiments, as a preferred embodiment, the data acquisition module is a module for acquiring overall power consumption and/or system processor utilization of the server.
Based on the foregoing embodiments, as a preferred embodiment, the data acquisition module is a module for out-of-band acquisition of the overall power consumption of the server, and/or in-band acquisition of the utilization of the system processor.
Based on the foregoing embodiments, as a preferred embodiment, the data acquisition module is a module for acquiring the overall power consumption of the server and/or the utilization of the system processor according to the set minimum acquisition frequency.
Based on the above embodiments, as a preferred embodiment, the data acquisition module further includes:
the acquisition detection unit is used for stopping acquiring the overall power consumption of the server if the utilization rate of the system processor is acquired; and if the utilization rate of the system processor is not acquired, acquiring the overall power consumption of the server, and carrying out normalization processing on the overall power consumption.
Based on the above embodiments, as a preferred embodiment, the acquisition detection unit includes:
the normalization unit is used for reading the power consumption of the whole machine; the power consumption of the whole machine comprises all acquired data in a set time length; determining a maximum power consumption value and a minimum power consumption value in the power consumption of the whole machine; and determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value.
Based on the above embodiments, as a preferred embodiment, the normalization processing unit includes:
the normalization processing unit is used for taking a larger value between the maximum power consumption value and the minimum power consumption value which is preset times as a power consumption large value and taking the minimum power consumption value as a power consumption small value; and taking the ratio of the difference value between the current power consumption and the small power consumption value to the difference value between the large power consumption value and the small power consumption value as normalized power consumption data.
Based on the above embodiments, as a preferred embodiment, the no-load detection module includes:
the average value calculation unit is used for calculating the average value of the index data;
the no-load calculation unit is used for substituting the average value into a no-load calculation formula to obtain a no-load calculation result;
the no-load detection unit is used for confirming that the server is in a no-load state if the no-load calculation result is not smaller than a preset value; and if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state.
Based on the above embodiments, as a preferred embodiment, the period analysis module includes:
a period change analysis unit configured to determine a period change of the index data; if the index data is changed in a strong period, counting the index data according to the strong period to obtain a period analysis result; if the index data is in weak periodic variation, predicting the index data by adopting a periodic prediction algorithm to obtain a periodic analysis result; and if the index data does not change periodically, predicting the index data by adopting a non-periodicity prediction algorithm to obtain a period analysis result.
Based on the above embodiments, as a preferred embodiment, the period change analysis unit includes:
the periodic variation detection unit is used for carrying out autocorrelation calculation on the index data and determining a periodic variation index; if the intra-period change index is larger than a first threshold value, confirming that the index data is in strong period change; if the intra-period change index is not greater than the first threshold and is greater than the second threshold, confirming that the index data is in weak period change; and if the periodic variation index is not greater than the second threshold value, confirming that the index data does not have periodic variation.
Based on the above-described embodiments, as a preferred embodiment, the period change detection unit includes:
the autocorrelation calculating unit is used for calculating the number of the acquisition points in the preset duration; squaring and summing index data acquired by all the points in the preset time length to obtain a first correlation value; multiplying the index data acquired in the preset time period by the index data acquired in the next preset time period in a one-to-one correspondence manner to obtain a second correlation value; calculating a correlation from the first correlation value and the second correlation value; and determining a period change index according to the maximum amplitude, the period change amplitude and the correlation.
Based on the above-described embodiments, as a preferred embodiment, the period change detection unit further includes:
and the data processing unit is used for carrying out fast Fourier transform on the index data to obtain the maximum amplitude and the periodical change amplitude.
Based on the above embodiments, as a preferred embodiment, the period change analysis unit includes:
and the strong period analysis unit is used for taking a time period corresponding to the time period when the average value of the index data in the strong period is minimum as a recommended time period of the operation and maintenance time window.
Based on the above embodiments, as a preferred embodiment, the period change analysis unit includes:
the weak period analysis unit is used for dividing the index data according to a set proportion to obtain training data and verification data; predicting the training data by adopting a plurality of period prediction algorithms; performing accuracy verification on each period prediction algorithm by using the verification data, and determining an optimal period prediction algorithm according to the symmetric average absolute percentage error; and predicting the index data by using the optimal period prediction algorithm to obtain a period analysis result.
Based on the above embodiments, as a preferred embodiment, the period change analysis unit includes:
The non-period analysis unit is used for dividing the index data according to a set proportion to obtain training data and verification data; predicting the training data by adopting a plurality of non-periodic prediction algorithms; performing accuracy verification on each of the non-periodic prediction algorithms by using the verification data, and determining an optimal non-periodic prediction algorithm according to the symmetric average absolute percentage error; and predicting the index data by using the optimal non-periodic prediction algorithm to obtain a periodic analysis result.
Based on the above embodiment, as a preferred embodiment, the weak period analysis unit is a unit for calculating a symmetric average absolute percentage error of each period prediction algorithm for performing accuracy verification on the verification data, and if the minimum value of the symmetric average absolute percentage error is smaller than the error preset value, taking the corresponding target period prediction algorithm as an optimal period prediction algorithm;
based on the foregoing embodiments, as a preferred embodiment, the mean value calculating unit is a unit for sorting the index data, and taking the index data before the fixed scale position to calculate the mean value.
Based on the above embodiment, as a preferred embodiment, further comprising:
And the proportion determining unit is used for determining a reliable data coefficient and determining the fixed proportion according to the reliable data coefficient.
The present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed, performs the steps provided by the above-described embodiments. The storage medium may include: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The present invention also provides a server, referring to fig. 4, and as shown in fig. 4, a structure diagram of a server provided in an embodiment of the present invention may include a processor 1410 and a memory 1420.
Processor 1410 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc., among others. The processor 1410 may be implemented in at least one hardware form of DSP (Digital Signal Processing ), FPGA (Field-Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array ). The processor 1410 may also include a main processor and a coprocessor, the main processor being a processor for processing data in an awake state, also referred to as a processor (Central Processing Unit, central processor); a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 1410 may integrate a GPU (Graphics Processing Unit, image processor) for rendering and drawing of content required to be displayed by the display screen. In some embodiments, the processor 1410 may also include an AI (Artificial Intelligence ) processor for processing computing operations related to machine learning.
Memory 1420 may include one or more computer-readable storage media, which may be non-transitory. Memory 1420 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In this embodiment, the memory 1420 is used at least to store a computer program 1421 that, when loaded and executed by the processor 1410, can implement relevant steps in the method performed by the server side as disclosed in any of the foregoing embodiments. In addition, the resources stored by memory 1420 may include an operating system 1422, data 1423, and the like, and the storage may be transient storage or permanent storage. The operating system 1422 may include Windows, linux, android, among other things.
In some embodiments, the server may further include a display 1430, an input-output interface 1440, a communication interface 1450, sensors 1460, a power supply 1470, and a communication bus 1480.
Of course, the structure of the server shown in fig. 4 is not limited to the server in the embodiment of the present invention, and the server may include more or less components than those shown in fig. 4, or may combine some components in practical applications.
In the description, each embodiment is described in a progressive manner, and each embodiment is mainly described by the differences from other embodiments, so that the same similar parts among the embodiments are mutually referred. The system provided by the embodiment is relatively simple to describe as it corresponds to the method provided by the embodiment, and the relevant points are referred to in the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to facilitate an understanding of the method of the present invention and its core ideas. It should be noted that it will be apparent to those skilled in the art that various modifications and adaptations of the invention can be made without departing from the principles of the invention and these modifications and adaptations are intended to be within the scope of the invention as defined in the following claims.
It should also be noted that in this specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Claims (21)
1. The operation and maintenance time window detection method of the server is characterized by comprising the following steps of:
collecting index data of a server;
determining whether the server is in an idle state according to the periodic variation state of the index data;
if the server is not in the idle state, performing periodic analysis on the index data to obtain a periodic analysis result; the period analysis is used for determining the period change of the index data and predicting by adopting a corresponding prediction algorithm based on the period change;
determining the operation and maintenance time window according to the period analysis result;
wherein the determining whether the server is in the idle state according to the periodically-changed state of the index data comprises:
calculating the average value of the index data;
substituting the average value into an idle load calculation formula to obtain an idle load calculation result;
if the idle load calculation result is not smaller than a preset value, confirming that the server is in an idle load state;
if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state;
the no-load calculation formula is as follows:
score=(0.1-Avg)*A+ (0.15-Avg99)*（1-A）；
wherein score is an idle calculation result, avg99 represents a mean value of the index data of the first 99% after the index data is sequenced, a is a coefficient, and Avg represents a mean value of the index data smaller than Avg 99.
2. The operation and maintenance time window detecting method according to claim 1, wherein the collecting the index data of the server includes:
and collecting the overall power consumption of the server and/or the utilization rate of a system processor.
3. The method for detecting an operation and maintenance time window according to claim 2, wherein the collecting the overall power consumption and/or the system processor utilization of the server comprises:
and collecting the overall power consumption of the server in an out-of-band manner and/or collecting the utilization rate of the system processor in an in-band manner.
4. The method for detecting an operation and maintenance time window according to claim 2, wherein the collecting the overall power consumption and/or the system processor utilization of the server comprises:
and acquiring the overall power consumption of the server and/or the utilization rate of a system processor according to the set minimum acquisition frequency.
5. The method for detecting an operation and maintenance time window according to claim 2, wherein when collecting the overall power consumption of the server and/or the utilization rate of a system processor, further comprises:
if the utilization rate of the system processor is acquired, stopping acquiring the overall power consumption of the server;
and if the utilization rate of the system processor is not acquired, acquiring the overall power consumption of the server, and carrying out normalization processing on the overall power consumption.
6. The operation and maintenance time window detection method according to claim 5, wherein the normalizing the power consumption of the whole machine comprises:
reading the power consumption of the whole machine; the power consumption of the whole machine comprises all acquired data in a set time length;
determining a maximum power consumption value and a minimum power consumption value in the power consumption of the whole machine;
and determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value.
7. The operation and maintenance time window detection method according to claim 6, wherein the determining normalized power consumption data based on the maximum power consumption value and the minimum power consumption value includes:
taking a larger value between the maximum power consumption value and the minimum power consumption value which is preset times as a large power consumption value, and taking the minimum power consumption value as a small power consumption value;
and taking the ratio of the difference value between the current power consumption and the small power consumption value to the difference value between the large power consumption value and the small power consumption value as normalized power consumption data.
8. The operation and maintenance time window detection method according to claim 1, wherein performing a period analysis on the index data to obtain a period analysis result includes:
determining a periodic variation of the index data;
If the index data is changed in a strong period, counting the index data according to the strong period to obtain a period analysis result;
if the index data is in weak periodic variation, predicting the index data by adopting a periodic prediction algorithm to obtain a periodic analysis result;
and if the index data does not change periodically, predicting the index data by adopting a non-periodicity prediction algorithm to obtain a period analysis result.
9. The operation and maintenance time window detection method according to claim 8, wherein the determining the periodic variation of the index data includes:
performing autocorrelation calculation on the index data, and determining a periodic variation index;
if the intra-period change index is larger than a first threshold value, confirming that the index data is in strong period change;
if the intra-period change index is not greater than the first threshold and is greater than the second threshold, confirming that the index data is in weak period change;
and if the periodic variation index is not greater than the second threshold value, confirming that the index data does not have periodic variation.
10. The method of claim 9, wherein the performing autocorrelation calculation on the index data, and determining the intra-period variation index comprises:
Calculating the number of the acquisition points in a preset time length;
squaring and summing index data acquired by all the points in the preset time length to obtain a first correlation value;
multiplying the index data acquired in the preset time period by the index data acquired in the next preset time period in a one-to-one correspondence manner to obtain a second correlation value;
calculating a correlation from the first correlation value and the second correlation value;
and determining a period change index according to the maximum amplitude, the period change amplitude and the correlation.
11. The method of claim 10, further comprising, prior to determining the intra-period variation index based on the maximum amplitude, the period variation amplitude, and the correlation:
and performing fast Fourier transform on the index data to obtain the maximum amplitude and the periodic variation amplitude.
12. The method for detecting an operation and maintenance time window according to claim 8, wherein the counting the index data according to the strong period to obtain the period analysis result comprises:
and taking a time period corresponding to the time period when the average value of the index data in the strong period is minimum as a recommended time period of the operation and maintenance time window.
13. The method for detecting an operation and maintenance time window according to claim 8, wherein predicting the index data by using a periodicity prediction algorithm to obtain a periodicity analysis result comprises:
dividing the index data according to a set proportion to obtain training data and verification data;
predicting the training data by adopting a plurality of period prediction algorithms;
performing accuracy verification on each period prediction algorithm by using the verification data, and determining an optimal period prediction algorithm according to the symmetric average absolute percentage error;
and predicting the index data by using the optimal period prediction algorithm to obtain a period analysis result.
14. The method for detecting an operation and maintenance time window according to claim 8, wherein predicting the index data by using a non-periodicity prediction algorithm to obtain a periodicity analysis result comprises:
dividing the index data according to a set proportion to obtain training data and verification data;
predicting the training data by adopting a plurality of non-periodic prediction algorithms;
performing accuracy verification on each of the non-periodic prediction algorithms by using the verification data, and determining an optimal non-periodic prediction algorithm according to the symmetric average absolute percentage error;
And predicting the index data by using the optimal non-periodic prediction algorithm to obtain a periodic analysis result.
15. The operation and maintenance time window detection method according to claim 13, wherein the performing accuracy verification on each of the cycle prediction algorithms using the verification data, determining an optimal cycle prediction algorithm based on a symmetric average absolute percentage error comprises:
and calculating the symmetrical average absolute percentage error of each period prediction algorithm for carrying out accuracy verification on the verification data, and taking the corresponding target period prediction algorithm as an optimal period prediction algorithm if the minimum value of the symmetrical average absolute percentage error is smaller than the error preset value.
16. The method of claim 15, wherein if the average absolute percentage error minimum is not less than the error preset value, further comprising:
and counting the index data according to the strong period to obtain a period analysis result.
17. The operation and maintenance time window detection method according to claim 1, wherein the calculating the mean value of the index data includes:
and sequencing the index data, and taking the index data before the fixed proportion position to calculate the average value.
18. The method for detecting an operation and maintenance time window according to claim 17, wherein before sorting the index data and taking the index data before the fixed scale position to calculate the average value, further comprising:
and determining a reliable data coefficient, and determining the fixed proportion according to the reliable data coefficient.
19. An operation and maintenance time window detection system of a server, comprising:
the data acquisition module is used for acquiring index data of the server;
the no-load detection module is used for determining whether the server is in an no-load state according to the periodic variation state of the index data;
the period analysis module is used for carrying out period analysis on the index data if the server is not in the idle state, so as to obtain a period analysis result; the period analysis is used for determining the period change of the index data and predicting by adopting a corresponding prediction algorithm based on the period change;
the time window detection module is used for determining the operation and maintenance time window according to the period analysis result;
wherein, no-load detection module includes:
the average value calculation unit is used for calculating the average value of the index data;
the no-load calculation unit is used for substituting the average value into a no-load calculation formula to obtain a no-load calculation result;
The no-load detection unit is used for confirming that the server is in a no-load state if the no-load calculation result is not smaller than a preset value; if the idle load calculation result is smaller than the preset value, confirming that the server is not in the idle load state;
the no-load calculation formula is as follows:
score=(0.1-Avg)*A+ (0.15-Avg99)*（1-A）；
wherein score is an idle calculation result, avg99 represents a mean value of the index data of the first 99% after the index data is sequenced, a is a coefficient, and Avg represents a mean value of the index data smaller than Avg 99.
20. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any of claims 1-18.
21. A server comprising a memory in which a computer program is stored and a processor which, when calling the computer program in the memory, implements the steps of the method according to any of claims 1-18.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311245830.6A CN116991448B (en) | 2023-09-26 | 2023-09-26 | Operation and maintenance time window detection method and system of server, storage medium and server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311245830.6A CN116991448B (en) | 2023-09-26 | 2023-09-26 | Operation and maintenance time window detection method and system of server, storage medium and server |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116991448A CN116991448A (en) | 2023-11-03 |
| CN116991448B true CN116991448B (en) | 2024-01-23 |
Family
ID=88525112
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311245830.6A Active CN116991448B (en) | 2023-09-26 | 2023-09-26 | Operation and maintenance time window detection method and system of server, storage medium and server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116991448B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114153685A (en) * | 2021-10-31 | 2022-03-08 | 济南浪潮数据技术有限公司 | Method, system, device and storage medium for estimating in-band system information of server |
| CN114201369A (en) * | 2022-02-17 | 2022-03-18 | 苏州浪潮智能科技有限公司 | Server cluster management method and device, electronic equipment and storage medium |
| CN114579554A (en) * | 2022-03-08 | 2022-06-03 | 京东科技信息技术有限公司 | Operation and maintenance time series data processing method and device, storage medium and electronic equipment |
| CN115659411A (en) * | 2022-09-26 | 2023-01-31 | 京东科技信息技术有限公司 | Method and device for data analysis |
-
2023
- 2023-09-26 CN CN202311245830.6A patent/CN116991448B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114153685A (en) * | 2021-10-31 | 2022-03-08 | 济南浪潮数据技术有限公司 | Method, system, device and storage medium for estimating in-band system information of server |
| CN114201369A (en) * | 2022-02-17 | 2022-03-18 | 苏州浪潮智能科技有限公司 | Server cluster management method and device, electronic equipment and storage medium |
| CN114579554A (en) * | 2022-03-08 | 2022-06-03 | 京东科技信息技术有限公司 | Operation and maintenance time series data processing method and device, storage medium and electronic equipment |
| CN115659411A (en) * | 2022-09-26 | 2023-01-31 | 京东科技信息技术有限公司 | Method and device for data analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116991448A (en) | 2023-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107871190A (en) | A kind of operational indicator monitoring method and device | |
| CN111813624B (en) | Robot execution time length estimation method based on time length analysis and related equipment thereof | |
| CN110444011B (en) | Traffic flow peak identification method and device, electronic equipment and storage medium | |
| CN111105786B (en) | Multi-sampling-rate voice recognition method, device, system and storage medium | |
| CN110764902A (en) | Virtual resource allocation method and device based on AI (Artificial Intelligence), computer equipment and storage medium | |
| CN111753875A (en) | Power information system operation trend analysis method and device and storage medium | |
| CN111381970A (en) | Cluster task resource allocation method and device, computer device and storage medium | |
| CN113673822B (en) | Elastic scheduling method and system | |
| CN116991448B (en) | Operation and maintenance time window detection method and system of server, storage medium and server | |
| CN113902260A (en) | Information prediction method, information prediction device, electronic equipment and medium | |
| CN115794532A (en) | Multi-index intelligent dynamic threshold monitoring method and system | |
| CN110928636A (en) | Virtual machine live migration method, device and equipment | |
| CN111768287A (en) | Period identification method, period identification device, server and readable storage medium | |
| US11636377B1 (en) | Artificial intelligence system incorporating automatic model updates based on change point detection using time series decomposing and clustering | |
| US11651271B1 (en) | Artificial intelligence system incorporating automatic model updates based on change point detection using likelihood ratios | |
| CN111080037A (en) | Short-term power load prediction method and device based on deep neural network | |
| CN112685390B (en) | Database instance management method and device and computing equipment | |
| CN115392591A (en) | Task processing method, device, equipment and storage medium | |
| CN114520773A (en) | Service request response method, device, server and storage medium | |
| CN112766535A (en) | Building load prediction method and system considering load curve characteristics | |
| CN112395167A (en) | Operation fault prediction method and device and electronic equipment | |
| US20190138931A1 (en) | Apparatus and method of introducing probability and uncertainty via order statistics to unsupervised data classification via clustering | |
| CN113391982B (en) | Monitoring data anomaly detection method, device and equipment | |
| CN110659655B (en) | Index classification method and device and computer readable storage medium | |
| CN115858291A (en) | System index detection method and device, electronic equipment and storage medium thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |