CN-122027464-A - Intelligent observation method and medium for equipment connection platform and electronic device

CN122027464ACN 122027464 ACN122027464 ACN 122027464ACN-122027464-A

Abstract

The application discloses an intelligent observation method, medium and electronic device for a device connection platform, and relates to the technical field of computers, wherein the method comprises the steps of obtaining a full-link buried point log; the method comprises the steps of determining corresponding evaluation indexes based on source channels of full-link buried point logs, expanding multidimensional attributes for the corresponding evaluation indexes by utilizing equipment identification, internet Protocol (IP) addresses and gateway nodes to associate external data sources to form corresponding equipment evaluation indexes with business contexts, and asynchronously calling an index insight intelligent body based on the equipment evaluation indexes to identify abnormal indexes deviating from normal periodicity or trend fluctuation according to the equipment evaluation indexes. The application solves the defects that the buried point logs are relatively scattered and the data volume is large, so that quick and accurate fault positioning cannot be realized, can identify abnormal indexes deviating from normal periodicity or trend fluctuation, avoid false alarm to business peak period, and simultaneously avoid alarm to normal growth trend, and realize quick and accurate positioning of the abnormal indexes.

Inventors

ZHANG YUNHAO
YANG MENG
YIN DESHUAI
HU BAICHUN

Assignees

青岛海尔科技有限公司
海尔优家智能科技（北京）有限公司
青岛海尔智能家电科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251225

Claims (10)

1. An intelligent observation method for a device connection platform is characterized by comprising the following steps: Acquiring a full-link buried point log; Determining corresponding evaluation indexes based on source channels of the full-link buried point logs, and expanding multidimensional attribute for the corresponding evaluation indexes by utilizing equipment identification, an Internet Protocol (IP) address and a gateway node associated external data source to form corresponding equipment evaluation indexes with business contexts; And asynchronously calling an index insight intelligent agent based on the equipment evaluation index to identify an abnormal index deviating from normal periodicity or trend fluctuation according to the equipment evaluation index.
2. The device connection platform intelligent observation method of claim 1, wherein determining a corresponding evaluation index based on a source channel of the full link buried point log comprises: determining a time-consuming index as a corresponding evaluation index according to the equipment gateway sending time and the message platform receiving time based on at least one of attribute reporting, version reporting and complete machine version of the source channel of the buried point log; Determining an online equipment number index as a corresponding evaluation index according to the online and offline logs of the equipment based on the source channel of the buried point log as an online and offline channel of the equipment; And determining a request number index and a connection state number index as corresponding evaluation indexes based on the source channel of the buried point log as a connection request channel, wherein the request number index is determined by requesting a device gateway by using a device gateway in advance, and the connection state number index is determined by responding to the device gateway by using the device gateway in advance.
3. The device connection platform intelligent observation method according to claim 2, wherein after determining an online device number indicator as the corresponding evaluation indicator, the method further comprises: Comparing the predicted online quantity of the current time window based on the prior prediction with the online equipment quantity index, and generating a dynamic fluctuation warning index when the actual online equipment quantity deviates from the predicted online quantity, wherein the predicted online quantity of the current time window is predicted based on a historical full-link buried point log; After determining the request number index and the connection state number index as the corresponding evaluation index, the method further includes: determining the increase rate of the request quantity in the request quantity index and the increase rate of the closing state quantity in the connection state index according to the request quantity index and the connection state quantity index; and when the increase rate of the request quantity is not matched with the increase rate of the closing state quantity, generating a predictive index of abnormal connection quality.
4. The intelligent observation method for the equipment connection platform according to claim 1, wherein the external data source comprises a distributed open source database hbase, an IP address location library IP2region and a gateway service library mysql, wherein the extending the multidimensional attribute for the corresponding evaluation index by using the equipment identifier, the internet protocol IP address and the gateway node to associate the external data source comprises: According to the evaluation index, combining the equipment identifier, the IP address and the gateway node, asynchronously initiating a parallel query request, and carrying out multidimensional expansion on the attribute of the corresponding evaluation index based on an asynchronous query return result, wherein the asynchronous initiation parallel query request is used for: Inquiring the hbase by using the equipment identifier to obtain equipment attribute information as a corresponding inquiry result; Inquiring the IP2region according to the IP address to obtain region information as a corresponding inquiry result; and inquiring the mysql according to the gateway node to obtain gateway group information as a corresponding inquiry result.
5. The device connection platform intelligent observation method of claim 4, wherein prior to asynchronously initiating a parallel query request, the method further comprises: Determining a priority strategy corresponding to dimension expansion according to the type and/or importance of the evaluation index, wherein the dimension expansion priority strategy is used for determining the priority based on the type and/or importance of the evaluation index and determining to initiate query requests to the hbase, the ip2region and/or the mysql in parallel according to the priority; Based on the selected dimension extended priority policy, a corresponding parallel query request is asynchronously initiated.
6. The device connection platform intelligent observation method according to claim 1, wherein after forming the device evaluation index corresponding to the service context, the method further comprises: Storing the equipment evaluation index into a time sequence database; Based on the device evaluation index, asynchronously invoking an index insight agent to identify an abnormal index deviating from a normal periodicity or trend fluctuation from the device evaluation index, comprising: And asynchronously calling an index insight intelligent agent based on the equipment evaluation index to pull the equipment evaluation index in the target time from the time sequence database, analyzing historical time sequence data of the equipment evaluation index by utilizing a long-short-term memory network or an isolated forest algorithm, and identifying an abnormal index deviating from normal periodicity or trend fluctuation.
7. The device connection platform intelligent observation method according to claim 1, wherein after asynchronously invoking an index insight agent based on the device evaluation index to identify an abnormal index deviating from a normal periodicity or a trend fluctuation from the device evaluation index, the method further comprises: based on the detection of the abnormal index, triggering association analysis on the multidimensional attribute to determine association degree of each dimensional attribute and the abnormal index, and positioning key dimensions causing the abnormality; Based on the detection of the abnormal index, carrying out path reasoning by utilizing a knowledge graph constructed based on equipment, a gateway and business services in advance, and determining a downstream service influenced by the abnormal index; And generating an insight report containing potential root causes and influence ranges according to the key dimension causing the abnormality and the downstream service influenced by the abnormality index.
8. The device connection platform intelligent observation method according to claim 1, wherein after forming the device evaluation index corresponding to the service context, the method further comprises: Storing the equipment evaluation index into a time sequence database; The method further comprises the steps of: Searching corresponding equipment evaluation indexes from the time sequence database by utilizing query sentences defined by a visualization platform in advance, wherein the visualization platform takes the time sequence database as a target data source in advance, and defines corresponding query sentences aiming at different types of visualization types; according to the found equipment evaluation index, analyzing multi-dimensional data, and generating or recommending a visual billboard template by combining a business target or a detected abnormal index; And according to the found equipment evaluation index, combining the visual billboard template to construct the visual billboard.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the program when run performs the method of any one of claims 1 to 8.
10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, the processor being arranged to perform the method of any of claims 1 to 8 by means of the computer program.

Description

Intelligent observation method and medium for equipment connection platform and electronic device Technical Field The present application relates to the field of computer technologies, and in particular, to an intelligent observation method for a device connection platform, a medium, and an electronic apparatus. Background With the rapid development of internet of things and industrial internet technologies, a device connection business system has become a core infrastructure in the fields of intelligent manufacturing, smart cities, intelligent energy sources and the like. The system realizes real-time acquisition, remote control and business coordination of data through wide access of mass equipment terminals, and the stability and reliability of the system are directly related to production efficiency and operation safety. In order to ensure the healthy operation of the system, the industry generally establishes a corresponding monitoring system to sense the operation state of the system in real time and alarm abnormality, thereby ensuring the continuity of the service. At present, monitoring of the running condition of a device connection service system is mainly realized by a technology based on a log embedded point, and the basic flow is that a data acquisition point (i.e. the embedded point) is preset at key nodes such as equipment, a gateway, a service link and the like and is used for recording key events such as heartbeat, online and offline states, data reporting, instruction response and the like of the equipment, collecting scattered log data in real time and converging the scattered log data to a centralized data platform (such as a data lake, a message queue or a log center), then cleaning, aggregating and analyzing the scattered log data through a data processing engine, finally displaying various core indexes (such as equipment online rate, message throughput, API response time and the like) through a visual instrument panel, and setting a static threshold alarm to realize macroscopic monitoring of the running state of the system. However, because the buried point logs are generated in different components and levels of the system, the data sources are naturally dispersed, and exponentially expand along with the increase of the equipment scale, so that the monitoring view angle is often isolated and local, systematic and integral cognition of the whole service system operation overall is difficult to form, the monitoring comprehensiveness is seriously insufficient, when the system is in a problem, the existing monitoring means can only provide alarming on the phenomenon level, the efficiency is low for the investigation of deep root causes, operation and maintenance personnel need to expend a great deal of effort to perform manual association analysis in massive and dispersed historical data, and comprehensively study and judge by combining the data trend and fluctuation, and the process is time-consuming and labor-consuming, and highly depends on the professional experience of personnel, so that quick and accurate fault positioning cannot be realized. Disclosure of Invention The application provides an intelligent observation method, medium and electronic device for an equipment connection platform, which are used for solving the defects that in the prior art, buried point logs are relatively scattered and the data volume is large so that quick and accurate fault positioning cannot be realized, identifying abnormal indexes deviating from normal periodicity or trend fluctuation, avoiding false alarm of service peak period, and simultaneously avoiding alarm of normal growth trend, so as to greatly filter invalid alarms and realize quick and accurate positioning of the abnormal indexes. The application provides an intelligent observation method for an equipment connection platform, which comprises the steps of obtaining a full-link buried point log, determining corresponding evaluation indexes based on a source channel of the full-link buried point log, expanding multidimensional attribute for the corresponding evaluation indexes by utilizing equipment identification, an Internet Protocol (IP) address and a gateway node associated external data source to form corresponding equipment evaluation indexes with business contexts, and asynchronously calling an index insight intelligent body based on the equipment evaluation indexes to identify abnormal indexes deviating from normal periodicity or trend fluctuation according to the equipment evaluation indexes. The intelligent observation method for the equipment connection platform provided by the application is based on a source channel of a full-link buried point log, and determines corresponding evaluation indexes, wherein the source channel of the buried point log is at least one of attribute reporting, version reporting and complete machine version, a time-consuming index is determined as the corresponding evaluation index according to equipmen