CN108345529A - A kind of system performance detection process and detection device - Google Patents
A kind of system performance detection process and detection device Download PDFInfo
- Publication number
- CN108345529A CN108345529A CN201810007610.2A CN201810007610A CN108345529A CN 108345529 A CN108345529 A CN 108345529A CN 201810007610 A CN201810007610 A CN 201810007610A CN 108345529 A CN108345529 A CN 108345529A
- Authority
- CN
- China
- Prior art keywords
- performance
- data
- index value
- performance data
- marketing system
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3452—Performance evaluation by statistical analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3409—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment for performance assessment
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of system performance detection process and detection devices, including：Predefine at least one performance point to be detected；For performance point to be detected described in each, the performance data corresponding with the performance point to be detected of acquisition at least two from external marketing system；Determine the corresponding data type of each collected described performance data；For data type described in each, each described performance data corresponding to the data type carries out data mart modeling, and obtains at least two index values；According to each described index value of acquisition, the marketing system performance is determined.This programme can improve the speed for finding marketing system abnormal performance.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of system performance detection process and detection device.
Background technology
Marketing system is that enterprise is that client creates value, and realization is exchanged with client's, and finally obtains income from sales and throwing
The thematic system for providing return, is to reach the most important guarantee of enterprise operation target.Therefore, entire running of the marketing system in enterprise
In have very important status.
Currently, most of the mode of detection marketing system performance is, after marketing system has handled certain data volume data
Detecting system performance again.But the data volume handled by marketing system is often larger, so that the time of processing data is opposite
Longer, when not being detected to the performance of marketing system also, marketing system may break down, so as to cause finding to market
The speed of system performance exception is slow.
Invention content
An embodiment of the present invention provides a kind of system performance detection process and detection devices, can improve discovery marketing system
The speed of abnormal performance.
In a first aspect, an embodiment of the present invention provides a kind of system performance detection process, including：
Predefine at least one performance point to be detected；
For performance point to be detected described in each, from external marketing system acquisition at least two with it is described to be detected
The corresponding performance data of performance point；
Determine the corresponding data type of each collected described performance data；
For data type described in each, each described performance data corresponding to the data type carries out data
Processing, and obtain at least two index values；
According to each described index value of acquisition, the marketing system performance is determined.
Preferably, described, acquisition at least two is corresponding with the performance point to be detected from external marketing system
After performance data, before the corresponding data type of each collected described performance data of the determination, further wrap
It includes：
For performance point to be detected described in each, at least one performance indicator is determined；
Initialize each described performance data of acquisition；
Described each described performance data corresponding to the data type carries out data mart modeling, and obtains at least one
Index value, including：
Filter redundancy performance data in each described performance data of acquisition；
For performance indicator described in each,
From each described performance data after filtering redundancy performance data, at least two and the performance indicator are determined
Corresponding specified performance data；
Determine the corresponding at least one index value of each performance data.
Preferably, data mart modeling is carried out in described each described performance data corresponding to the data type, and obtained
After at least two index values, in described each described index value according to acquisition, determine the marketing system performance it
Before, further comprise：
Determine the corresponding at least one feature of each index value；
Described each described index value according to acquisition determines the marketing system performance, including：
In conjunction with each determining feature and each described index value, performance graph is generated；
According to the performance graph of generation, the marketing system performance is determined.
Preferably, it is directed to each described data type described, at least two corresponding to each index value are special
After sign, further comprise：
For each the described feature determined, by each described index value and by each history gathered in advance
The corresponding at least one index value that prestores of data carries out tagsort according to the feature and forms at least one feature class；
It is for feature class described in each, each corresponding described index value of the feature class is described pre- with each
It deposits index value to be compared, determines the marketing system performance with the presence or absence of abnormal.
Preferably, described according to the performance point to be detected, acquisition at least two and institute from external marketing system
Before stating the corresponding performance data of performance point to be detected, further comprise：
Pre-set at least one server；
According to the corresponding attribute of server described in each and preset node rule, from server described in each
Middle determining given server is as host node, wherein the attribute, including cpu performance and/or IP address；
By the host node, the corresponding Detection task of each remaining described server is determined respectively, wherein described
It is any one or more in Detection task, including performance data collection, performance data processing and performance data analysis；
The acquisition at least two from external marketing system performance data corresponding with the performance point to be detected,
Including：
For each remaining described server, executed according to the corresponding Detection task described from marketing system
The performance data corresponding with the performance point to be detected of acquisition at least two in system.
Second aspect, an embodiment of the present invention provides a kind of system performance detection devices, including：
Setting unit, for predefining at least one performance point to be detected；Determine each collected described performance
The corresponding data type of data；
Collecting unit determines each described performance point to be detected for being directed to shown setting unit, from external marketing
The performance data corresponding with the performance point to be detected of acquisition at least two in system；
Processing unit, each described data type for being determined for shown setting unit, to the data type
The performance data of each corresponding collecting unit acquisition carries out data mart modeling, and obtains at least two index values；
According to each described index value of acquisition, the marketing system performance is determined.
Preferably, the setting unit determines that at least one performance refers to for being directed to each described performance point to be detected
Mark；
The processing unit, each described performance data for initializing acquisition；Described in each for filtering acquisition
Redundancy performance data in performance data；For performance indicator described in each, each institute after filtering redundancy performance data
It states in performance data, determines at least two specified performance data corresponding with the performance indicator；Determine each performance
The corresponding at least one index value of data.
Preferably, the processing unit, for determining the corresponding at least one feature of each index value；In conjunction with determination
Each described feature and each described index value, generate performance graph；According to the performance graph of generation, institute is determined
State marketing system performance.
Preferably, the processing unit is further used for for each the described feature determined, described in each
Index value carries out special with by the corresponding at least one index value that prestores of each historical data gathered in advance according to the feature
Sign classification forms at least one feature class；For feature class described in each, by each corresponding described finger of the feature class
Scale value and the index value that prestores described in each are compared, and determine the marketing system performance with the presence or absence of abnormal.
Preferably, the setting unit is further used for pre-setting at least one server；According to clothes described in each
The business corresponding attribute of device and preset node rule determine given server as main section from server described in each
Point, wherein the attribute, including cpu performance and/or IP address；By the host node, each remaining institute is determined respectively
State the corresponding Detection task of server, wherein the Detection task, including performance data collection, performance data processing and performance
It is any one or more in data analysis；
The collecting unit is further used for being directed to each remaining described server, according to the corresponding detection
Described in task execution from the marketing system performance data corresponding with the performance point to be detected of acquisition at least two.
In embodiments of the present invention, from external marketing system collecting performance data (for example, server log, number
According to library daily record) when, it needs to be adopted according to predetermined performance point (for example, server performance, database performance) to be detected
Collection, then determine the data type of performance data, i.e., classification and data working process are carried out to performance data, and obtain and refer to accordingly
Scale value, the index value by acquisition is that can determine the performance of marketing system, without waiting until that it is a certain amount of that marketing system has been handled
The performance of marketing system is determined after data again, thus can be broken down to avoid marketing system when handling data and can not be timely
It was found that so as to improve the speed for finding marketing system abnormal performance.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is the present invention
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.
Fig. 1 is a kind of flow chart for system performance detection process that one embodiment of the invention provides；
Fig. 2 is a kind of structural schematic diagram for system performance detection device that one embodiment of the invention provides；
Fig. 3 is the structural schematic diagram for another system performance detection device that one embodiment of the invention provides.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments, based on the embodiments of the present invention, those of ordinary skill in the art
The every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
As shown in Figure 1, an embodiment of the present invention provides a kind of system performance detection process and detection devices, including：
Step 101：Predefine at least one performance point to be detected；
Step 102：For performance point to be detected described in each, acquisition at least two and institute from external marketing system
State the corresponding performance data of performance point to be detected；
Step 103：Determine the corresponding data type of each collected described performance data；
Step 104：For data type described in each, each described performance data corresponding to the data type
Data mart modeling is carried out, and obtains at least two index values；
Step 105：According to each described index value of acquisition, the marketing system performance is determined.
In embodiments of the present invention, from external marketing system collecting performance data (for example, server log, number
According to library daily record) when, it needs to be adopted according to predetermined performance point (for example, server performance, database performance) to be detected
Collection, then determine the data type of performance data, i.e., classification and data working process are carried out to performance data, and obtain and refer to accordingly
Scale value, the index value by acquisition is that can determine the performance of marketing system, without waiting until that it is a certain amount of that marketing system has been handled
The performance of marketing system is determined after data again, thus can be broken down to avoid marketing system when handling data and can not be timely
It was found that so as to improve the speed for finding marketing system abnormal performance.
In an embodiment of the present invention, at least two and the property to be detected are acquired from external marketing system described
After corresponding performance data capable of being put, the corresponding data type of each collected described performance data of the determination it
Before, further comprise：
For performance point to be detected described in each, at least one performance indicator is determined；
Initialize each described performance data of acquisition；
Described each described performance data corresponding to the data type carries out data mart modeling, and obtains at least one
Index value, including：
Filter redundancy performance data in each described performance data of acquisition；
For performance indicator described in each,
From each described performance data after filtering redundancy performance data, at least two and the performance indicator are determined
Corresponding specified performance data；
Determine the corresponding at least one index value of each performance data.It is acquired from external marketing system described
After at least two performance datas corresponding with the performance point to be detected, in each collected described property of the determination
Before the corresponding data type of energy data, further comprise：
For performance point to be detected described in each, at least one performance indicator is determined；
Initialize each described performance data of acquisition；
Described each described performance data corresponding to the data type carries out data mart modeling, and obtains at least one
Index value, including：
Filter redundancy performance data in each described performance data of acquisition；
For performance indicator described in each,
From each described performance data after filtering redundancy performance data, at least two and the performance indicator are determined
Corresponding specified performance data；
Determine the corresponding at least one index value of each performance data.
In embodiments of the present invention, it after collecting performance data, needs to carry out initialization process to each performance data,
So that difficulty when reducing server process performance data, and redundancy, imperfect and wrong performance number in strainability data
According to, then unify the format of each performance data, and according to the performance point of various dimensions, determine the corresponding multiple indexs of performance data
Value, you can the marketing system performance outside determining.
Wherein, the performance data after memory data mart modeling can be also saved in pre-set Mysql databases different
Tables of data in, so as to determine index value to be follow-up and determine that marketing system performance provides data source, and performance data is preserved
Into different tables of data, you can increase the speed of process performance data, and the flexibility of process performance data can be improved.
In an embodiment of the present invention, in described each described performance data corresponding to the data type into line number
According to processing, and after at least two index values of acquisition, in described each described index value according to acquisition, determine the marketing
Before system performance, further comprise：
Determine the corresponding at least one feature of each index value；
Described each described index value according to acquisition determines the marketing system performance, including：
In conjunction with each determining feature and each described index value, performance graph is generated；
According to the performance graph of generation, the performance of the marketing system is determined.
In embodiments of the present invention, index value (for example, index value is the occupancy of some time point CPU) is being determined
Afterwards, its dispersion is matched by machine learning K-Means clustering algorithms, determines the corresponding feature of each index value, (for example,
Change over time, CPU usage increases, reduces), then index value that each feature and each performance point to be monitored are determined into
Row processing, you can generate the performance graph of various dimensions, then show user, user can be according to performance the performance graph of generation
Chart determines marketing system performance, to realize real-time informing and the early warning of marketing system performance.
In an embodiment of the present invention, it is directed to each described data type described, corresponding to each index value
At least two features after, further comprise：
For each the described feature determined, by each described index value and by each history gathered in advance
The corresponding at least one index value that prestores of data carries out tagsort according to the feature and forms at least one feature class；
It is for feature class described in each, each corresponding described index value of the feature class is described pre- with each
It deposits index value to be compared, determines the marketing system performance with the presence or absence of abnormal.
In embodiments of the present invention, by the basic classification model under machine decision Tree algorithms, to each spy determined
Sign carries out classification and forms each feature class, and combines historical data gathered in advance, you can marketing system performance characteristic is summarized, and
The principal element for influencing system performance can be excavated with depth, to realize the purpose for finding potential risk existing for marketing system.
In an embodiment of the present invention, it is acquired from external marketing system according to the performance point to be detected described
Before at least two performance datas corresponding with the performance point to be detected, further comprise：
Pre-set at least one server；
According to the corresponding attribute of server described in each and preset node rule, from server described in each
Middle determining given server is as host node, wherein the attribute, including cpu performance and/or IP address；
By the host node, the corresponding Detection task of each remaining described server is determined respectively, wherein described
It is any one or more in Detection task, including performance data collection, performance data processing and performance data analysis；
The acquisition at least two from external marketing system performance data corresponding with the performance point to be detected,
Including：
For each remaining described server, executed according to the corresponding Detection task described from marketing system
The performance data corresponding with the performance point to be detected of acquisition at least two in system.
In embodiments of the present invention, before the marketing system performance data outside acquisition, need first to dispose being based on
The multi node server of Hadoop environment, and Hadoop environment is deployed on cheap server (PC), it can not only provide
High-throughput accesses the performance data of off-line evolution system, can also reduce the cost of process performance data.Again from multiple clothes
It, can be according to the corresponding attribute (example of each server when being engaged in determining that a server is host node in device, and determining host node
Such as, CPU, IP) it determines, and data flow allotment, task scheduling etc. are carried out to remaining server by host node, it can both improve
Determine the speed of off-line evolution system performance, additionally it is possible to Single Point of Faliure is effectively prevent, to achieve the purpose that improve redundancy ability.
Wherein, Detection task can be from marketing system collecting performance data, determine the number of collected performance data
According to type, obtains index value, determines that any one or more and host node in marketing system performance can be in remaining clothes
Installation and deployment Spark, HDFS, Hive distributed in device of being engaged in calculates, so that the performance data processing for marketing system provides fortune
Row environment.
In order to more clearly illustrate technical scheme of the present invention and advantage, below using performance point to be detected as server
For energy, a kind of system performance detection process provided in an embodiment of the present invention is described in detail, this method may include with
Lower step：
Step 201：At least two performance datas corresponding with server performance are acquired from external marketing system.
Specifically, it before the performance data of acquisition off-line evolution system, needs first to dispose based on the more of Hadoop environment
Node server, and host node is selected from multiple servers, it is the global allotment of remaining server progress by host node, and
According to the operational indicator in the actual demand of user, to determine the corresponding performance point to be detected of marketing system, so that according to performance
Point collecting performance data from external marketing system.
For example, server f, server w and server q are pre-set；
The corresponding attributes of server f be IP address be 1, the corresponding attributes of server w be IP address be 2, server q is corresponded to
Attribute be IP address be 3, host node is determined according to the sequencing that default rule is IP address, is determined based on server f
Node；
Management node is installed in server f, and by server f be server w and server q install respectively Spark,
HDFS、Hive；
Server w and server q is server performance according to predetermined performance point, is adopted from external marketing system
Collecting performance data is：
12 days 14 December in 2017：10：30%, 2017 on December 12,14 of EMS memory occupation：20：EMS memory occupation 50%；
12 days 14 December in 2017：10：60%, 2017 on December 12,14 of CPU usage:50：CPU usage 65%,
12 days 15 December in 2017：10：CPUzhanyonglv0.8,12 days 15 December in 2017：50：CPU usage 85%.
Step 202：Initialize each performance data of acquisition.
Specifically, follow-up using the performance data acquired for convenience, the performance data for needing initialization to acquire.
Step 203：Filter redundancy performance data in each performance data of acquisition.
Specifically, after the performance data of initialization acquisition, redundancy in strainability data, imperfect, mistake is needed
Performance data, so as to can be used in subsequent processing by filtered performance data.
For example, filtering redundancy performance data is 2017.12.14.14.20CPU occupancies 65%, by performance data
12 days 15 December in 2017：10：The format conversion of CPUzhanyonglv0.8 is 12 days 15 December in 2017：10：CPU usage
80%.
Step 204：Determine the corresponding at least one performance indicator of server performance.
Specifically, it when marketing system is run, needs to carry out follow-up investigation to marketing system, need for marketing system
Each performance point (for example, server performance, database performance) to be monitored splits into multiple performance indicators (for example, CPU is provided
Source, EMS memory occupation, capacity prediction, handling capacity, response time, CPU pressure, disk occupy), so as to be determined by performance indicator
The operating condition of marketing system could be subsequently to provide reference, and adjusting early is to avoid failure problems.
For example, determine that performance point to be monitored is the CPU usage in server performance according to operational indicator.
Step 205：For each performance indicator, from each performance data after filtering redundancy performance data, really
Fixed at least two specified performance data corresponding with performance indicator.
Specifically, each performance indicator can correspond to multiple specified performance data in the performance data of acquisition, further according to
Rule between each specified performance data, determines the operating condition of marketing system.
For example, determine with performance indicator to be that the corresponding specified performance data of CPU usage are：
12 days 14 December in 2017：10：60%, 2017 on December 12,14 of CPU usage:50：CPU usage 65%,
12 days 15 December in 2017：10：80%, 2017 on December 12,15 of CPU usage：50：CPU usage 85%.
Step 206：Determine the corresponding at least one index value of each specified performance data.
Specifically, each specified performance data are combined with default rule, you can determine corresponding at least one
Index value.
For example, as unit of hour, determine that specified performance data are 12 days 14 December in 2017：10：CPU usage
60% and 12 days 14 December in 2017:50：65% corresponding index value of CPU usage is CPU usage 63%.
Determine that specified performance data are 12 days 15 December in 2017：10：80%, 2017 on December 12,15 of CPU usage：
50：85% corresponding index value of CPU usage is CPU usage 83%.
Step 207：Determine the corresponding at least one feature of each index value.
Specifically, process and its study K-Means clustering algorithms match the dispersion of each index value, you can determine and correspond to
Feature.
For example, it determines that index value is that CPU usage 63% is corresponding and is characterized as being less than 70%, determines that index value is
CPU usage 83% is corresponding to be characterized as being more than 80%.
Step 208：In conjunction with each determining feature and each index value, performance graph is generated.
Specifically, each index value is ranked up, corresponding performance graph is produced in conjunction with corresponding feature.
For example, marketing system can be drawn out as time index value increases again according to each index value of time-sequencing
The performance graph of CPU.
Step 209：According to the performance graph of generation, the performance of marketing system is determined.
Specifically, after generating performance graph, show user, user can basis in a manner of user interface interaction UI
Information in performance graph determines the performance of marketing system, to realize real-time informing and the early warning of marketing system performance, also,
System performance character is summarized, depth, which is excavated, influences system based on the basic classification model under decision Tree algorithms by machine learning again
The principal element for performance of uniting.Potential risk existing for timely discovery system.
As shown in figure 3, an embodiment of the present invention provides a kind of system performance detection devices, including：
Setting unit 301, for predefining at least one performance point to be detected；Determine that collecting unit is collected each
The corresponding data type of a performance data；
Collecting unit 302, for determining each described performance point to be detected for shown setting unit 301, from outside
Marketing system in the performance data corresponding with the performance point to be detected of acquisition at least two；
Processing unit 303, each described data type for being determined for shown setting unit 301, to the number
Data mart modeling is carried out according to the performance data of corresponding each described collecting unit 302 acquisition of type, and obtains at least two
A index value；According to each described index value of acquisition, the marketing system performance is determined.
In embodiments of the present invention, collecting unit from external marketing system collecting performance data (for example, server
Daily record, database journal) when, it needs according to the predetermined performance point to be detected of setting unit (for example, server performance, number
According to library performance) it is acquired, processing unit determines the data type of performance data by setting unit again, i.e., is adopted to collecting unit
The performance data of collection carries out classification and data working process, and obtains corresponding index value, can be true by the index value of acquisition
The performance for determining marketing system, without determining the performance of marketing system again after marketing system has handled a certain amount of data,
Therefore to avoid marketing system breaks down when handling data and can not find in time marketing system can be found so as to improve
The speed of system abnormal performance.
In an embodiment of the present invention, the setting unit determines extremely for being directed to each described performance point to be detected
A few performance indicator；
The processing unit, each described performance data for initializing acquisition；Described in each for filtering acquisition
Redundancy performance data in performance data；For performance indicator described in each, each institute after filtering redundancy performance data
It states in performance data, determines at least two specified performance data corresponding with the performance indicator；Determine each performance
The corresponding at least one index value of data.
In an embodiment of the present invention, the processing unit, for determining that each index value is corresponding at least one
Feature；In conjunction with each determining feature and each described index value, performance graph is generated；According to the property of generation
Energy chart, determines the marketing system performance.
In an embodiment of the present invention, the processing unit is further used for for each the described feature determined,
By each described index value with by the corresponding at least one index value that prestores of each historical data gathered in advance, according to institute
It states feature progress tagsort and forms at least one feature class；It is for feature class described in each, the feature class is corresponding
Each described index value and the index value that prestores described in each are compared, and determine the marketing system performance with the presence or absence of different
Often.
In an embodiment of the present invention, the setting unit is further used for pre-setting at least one server；According to
Each corresponding attribute of server and preset node rule determine specified clothes from server described in each
Device be engaged in as host node, wherein the attribute, including cpu performance and/or IP address；By the host node, determine respectively surplus
The corresponding Detection task of each remaining server, wherein the Detection task, including performance data collection, performance number
According to any one or more in processing and performance data analysis；
The collecting unit is further used for being directed to each remaining described server, according to the corresponding detection
Described in task execution from the marketing system performance data corresponding with the performance point to be detected of acquisition at least two.
The each embodiment of the present invention at least has the advantages that：
1, in an embodiment of the present invention, from external marketing system collecting performance data (for example, server day
Will, database journal) when, need according to predetermined performance point (for example, server performance, database performance) to be detected into
Row acquisition, then determine the data type of performance data, i.e., classification and data working process are carried out to performance data, and obtain corresponding
Index value, be that can determine the performance of marketing system by the index value of acquisition, without waiting until that marketing system has been handled centainly
The performance of marketing system is determined after the data of amount again, thus can be broken down to avoid marketing system when handling data and can not
It finds in time, so as to improve the speed for finding marketing system abnormal performance.
2, in an embodiment of the present invention, it after collecting performance data, needs to initialize each performance data
Processing, so that difficulty when reducing server process performance data, and redundancy, imperfect and wrong property in strainability data
Energy data, then unify the format of each performance data, and according to the performance point of various dimensions, determine the corresponding multiple fingers of performance data
Scale value, you can the marketing system performance outside determining.
3, in an embodiment of the present invention, determining index value (for example, index value is the occupancy of some time point CPU
Rate) after, its dispersion is matched by machine learning K-Means clustering algorithms, determines the corresponding feature of each index value, (example
Such as, change over time, CPU usage increases, reduces), then to index that each feature and each performance point to be monitored are determined
Value is processed, you can is generated the performance graph of various dimensions, then is showed user, user can basis the performance graph of generation
Performance graph determines marketing system performance, to realize real-time informing and the early warning of marketing system performance.
4, in an embodiment of the present invention, each to what is determined by the basic classification model under machine decision Tree algorithms
A feature carries out classification and forms each feature class, and combines historical data gathered in advance, you can it is special to summarize marketing system performance
Sign, and the principal element for influencing system performance can be excavated with depth, find potential risk existing for marketing system to realize
Purpose.
5, in an embodiment of the present invention, before the marketing system performance data outside acquisition, need first to dispose being based on
The multi node server of Hadoop environment, and Hadoop environment is deployed on cheap server (PC), it can not only provide
High-throughput accesses the performance data of off-line evolution system, can also reduce the cost of process performance data.Again from multiple clothes
It, can be according to the corresponding attribute (example of each server when being engaged in determining that a server is host node in device, and determining host node
Such as, CPU, IP) it determines, and data flow allotment, task scheduling etc. are carried out to remaining server by host node, it can both improve
Determine the speed of off-line evolution system performance, additionally it is possible to Single Point of Faliure is effectively prevent, to achieve the purpose that improve redundancy ability.
It should be noted that herein, such as first and second etc relational terms are used merely to an entity
Or operation is distinguished with another entity or operation, is existed without necessarily requiring or implying between these entities or operation
Any actual relationship or order.Moreover, the terms "include", "comprise" or its any other variant be intended to it is non-
It is exclusive to include, so that the process, method, article or equipment including a series of elements includes not only those elements,
But also include other elements that are not explicitly listed, or further include solid by this process, method, article or equipment
Some elements.In the absence of more restrictions, the element limited by sentence " including a 〃 〃 ", it is not excluded that
There is also other identical factors in the process, method, article or apparatus that includes the element.
Finally, it should be noted that：The foregoing is merely presently preferred embodiments of the present invention, is merely to illustrate the skill of the present invention
Art scheme, is not intended to limit the scope of the present invention.Any modification for being made all within the spirits and principles of the present invention,
Equivalent replacement, improvement etc., are included within the scope of protection of the present invention.
Claims (10)
1. a kind of system performance detection process, which is characterized in that including：
Predefine at least one performance point to be detected；
For performance point to be detected described in each, acquisition at least two and the performance to be detected from external marketing system
The corresponding performance data of point；
Determine the corresponding data type of each collected described performance data；
For data type described in each, each described performance data corresponding to the data type carries out data and adds
Work, and obtain at least two index values；
According to each described index value of acquisition, the marketing system performance is determined.
2. detection method according to claim 1, which is characterized in that
The acquisition at least two from external marketing system performance data corresponding with the performance point to be detected it
Afterwards, before the corresponding data type of each collected described performance data of the determination, further comprise：
For performance point to be detected described in each, at least one performance indicator is determined；
Initialize each described performance data of acquisition；
Described each described performance data corresponding to the data type carries out data mart modeling, and obtains at least one index
Value, including：
Filter redundancy performance data in each described performance data of acquisition；
For performance indicator described in each,
From each described performance data after filtering redundancy performance data, determine that at least two is opposite with the performance indicator
The specified performance data answered；
Determine the corresponding at least one index value of each performance data.
3. detection method according to claim 1, which is characterized in that
Data mart modeling is carried out in described each described performance data corresponding to the data type, and obtains at least two and refers to
After scale value, in described each described index value according to acquisition, before determining the marketing system performance, further wrap
It includes：
Determine the corresponding at least one feature of each index value；
Described each described index value according to acquisition determines the marketing system performance, including：
In conjunction with each determining feature and each described index value, performance graph is generated；
According to the performance graph of generation, the marketing system performance is determined.
4. detection method according to claim 3, which is characterized in that
It is directed to each described data type described, after at least two features corresponding to each index value, into one
Step includes：
For each the described feature determined, by each described index value and by each historical data gathered in advance
Corresponding at least one index value that prestores carries out tagsort according to the feature and forms at least one feature class；
For feature class described in each, by the finger that prestores described in each corresponding described index value of the feature class and each
Scale value is compared, and determines the marketing system performance with the presence or absence of abnormal.
5. according to any detection method in Claims 1-4, which is characterized in that
Described according to the performance point to be detected, acquisition at least two and the performance to be detected from external marketing system
Before the corresponding performance data of point, further comprise：
Pre-set at least one server；
According to the corresponding attribute of server described in each and preset node rule, from server described in each really
Given server is determined as host node, wherein the attribute, including cpu performance and/or IP address；
By the host node, the corresponding Detection task of each remaining described server is determined respectively, wherein the detection
It is any one or more in task, including performance data collection, performance data processing and performance data analysis；
The acquisition at least two from external marketing system performance data corresponding with the performance point to be detected, is wrapped
It includes：
For each remaining described server, executed according to the corresponding Detection task described from the marketing system
The performance data corresponding with the performance point to be detected of acquisition at least two.
6. a kind of system performance detection device, which is characterized in that including：
Setting unit, for predefining at least one performance point to be detected；Determine each collected described performance data
Corresponding data type；
Collecting unit determines each described performance point to be detected for being directed to shown setting unit, from external marketing system
At least two performance data corresponding with the performance point to be detected of middle acquisition；
Processing unit, each described data type for being determined for shown setting unit correspond to the data type
The performance data of each described collecting unit acquisition carry out data mart modeling, and obtain at least two index values；According to
Each the described index value obtained, determines the marketing system performance.
7. detection device according to claim 6, which is characterized in that
The setting unit determines at least one performance indicator for being directed to each described performance point to be detected；
The processing unit, each described performance data for initializing acquisition；Filter each described performance of acquisition
Redundancy performance data in data；For performance indicator described in each, each described property after filtering redundancy performance data
In energy data, at least two specified performance data corresponding with the performance indicator are determined；Determine each performance data
Corresponding at least one index value.
8. detection device according to claim 6, which is characterized in that
The processing unit, for determining the corresponding at least one feature of each index value；In conjunction with each determining institute
Feature and each described index value are stated, performance graph is generated；According to the performance graph of generation, the marketing system is determined
Performance.
9. detection device according to claim 6, which is characterized in that
The processing unit is further used for, for each the described feature determined, by each described index value and inciting somebody to action
The corresponding at least one index value that prestores of each historical data gathered in advance carries out tagsort according to the feature and is formed
At least one feature class；For feature class described in each, by each corresponding described index value of the feature class and each
A index value that prestores is compared, and determines the marketing system performance with the presence or absence of abnormal.
10. according to any detection device in claim 6 to 9, which is characterized in that
The setting unit is further used for pre-setting at least one server；It is right respectively according to server described in each
The attribute answered and preset node rule determine given server as host node, wherein institute from server described in each
State attribute, including cpu performance and/or IP address；By the host node, each remaining described server pair is determined respectively
The Detection task answered, wherein in the Detection task, including performance data collection, performance data processing and performance data analysis
It is any one or more；
The collecting unit is further used for being directed to each remaining described server, according to the corresponding Detection task
Execute the acquisition at least two from marketing system performance data corresponding with the performance point to be detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810007610.2A CN108345529A (en) | 2018-01-04 | 2018-01-04 | A kind of system performance detection process and detection device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810007610.2A CN108345529A (en) | 2018-01-04 | 2018-01-04 | A kind of system performance detection process and detection device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108345529A true CN108345529A (en) | 2018-07-31 |
Family
ID=62960586
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810007610.2A Pending CN108345529A (en) | 2018-01-04 | 2018-01-04 | A kind of system performance detection process and detection device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108345529A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674009A (en) * | 2019-09-10 | 2020-01-10 | 平安普惠企业管理有限公司 | Application server performance monitoring method and device, storage medium and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102664961A (en) * | 2012-05-04 | 2012-09-12 | 北京邮电大学 | Method for anomaly detection in MapReduce environment |
| CN103236000A (en) * | 2013-04-16 | 2013-08-07 | 中国南方电网有限责任公司 | Index data processing system |
| CN103246569A (en) * | 2013-05-20 | 2013-08-14 | 浪潮(北京)电子信息产业有限公司 | Method and device for representing high-performance calculation application characteristics |
| US20150278079A1 (en) * | 2014-03-31 | 2015-10-01 | International Business Machines Corporation | Creating software performance testing environment based on virtual machine |
| CN105183627A (en) * | 2015-10-20 | 2015-12-23 | 浪潮(北京)电子信息产业有限公司 | Server performance prediction method and system |
| CN106789270A (en) * | 2016-12-27 | 2017-05-31 | 浪潮软件集团有限公司 | Method and system for realizing centralized operation and maintenance management of information system |
-
2018
- 2018-01-04 CN CN201810007610.2A patent/CN108345529A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102664961A (en) * | 2012-05-04 | 2012-09-12 | 北京邮电大学 | Method for anomaly detection in MapReduce environment |
| CN103236000A (en) * | 2013-04-16 | 2013-08-07 | 中国南方电网有限责任公司 | Index data processing system |
| CN103246569A (en) * | 2013-05-20 | 2013-08-14 | 浪潮(北京)电子信息产业有限公司 | Method and device for representing high-performance calculation application characteristics |
| US20150278079A1 (en) * | 2014-03-31 | 2015-10-01 | International Business Machines Corporation | Creating software performance testing environment based on virtual machine |
| US20160378643A1 (en) * | 2014-03-31 | 2016-12-29 | International Business Machines Corporation | Creating software performance testing environment based on virtual machine |
| CN105183627A (en) * | 2015-10-20 | 2015-12-23 | 浪潮(北京)电子信息产业有限公司 | Server performance prediction method and system |
| CN106789270A (en) * | 2016-12-27 | 2017-05-31 | 浪潮软件集团有限公司 | Method and system for realizing centralized operation and maintenance management of information system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110674009A (en) * | 2019-09-10 | 2020-01-10 | 平安普惠企业管理有限公司 | Application server performance monitoring method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11405301B1 (en) | Service analyzer interface with composite machine scores | |
| CN110321371B (en) | Log data anomaly detection method, device, terminal and medium | |
| US11657309B2 (en) | Behavior analysis and visualization for a computer infrastructure | |
| US10649838B2 (en) | Automatic correlation of dynamic system events within computing devices | |
| US11586972B2 (en) | Tool-specific alerting rules based on abnormal and normal patterns obtained from history logs | |
| US10592308B2 (en) | Aggregation based event identification | |
| US10002144B2 (en) | Identification of distinguishing compound features extracted from real time data streams | |
| US11138058B2 (en) | Hierarchical fault determination in an application performance management system | |
| US8924328B1 (en) | Predictive models for configuration management of data storage systems | |
| EP3031216A1 (en) | Dynamic collection analysis and reporting of telemetry data | |
| US11093319B2 (en) | Automated recovery of webpage functionality | |
| CN104903866A (en) | Management system and method for assisting event root cause analysis | |
| US10713140B2 (en) | Identifying latent states of machines based on machine logs | |
| CN105404581A (en) | Database evaluation method and device | |
| US20220036224A1 (en) | Determination of storage configuration for enterprise distributed environment | |
| US8782341B1 (en) | Embedded advisory framework for storage configuration management | |
| Sukhija et al. | Event management and monitoring framework for HPC environments using ServiceNow and Prometheus | |
| US11153183B2 (en) | Compacted messaging for application performance management system | |
| CN108345529A (en) | A kind of system performance detection process and detection device | |
| US11334410B1 (en) | Determining aberrant members of a homogenous cluster of systems using external monitors | |
| US10848371B2 (en) | User interface for an application performance management system | |
| JP6845657B2 (en) | Management server, management method and its program | |
| US20210092159A1 (en) | System for the prioritization and dynamic presentation of digital content | |
| EP2915059B1 (en) | Analyzing data with computer vision | |
| CN109074269B (en) | Determining equipment wear rate |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180731 |