Device monitoring alarm processing method, device, medium and equipment based on zabbix

By acquiring monitoring data through zabbix-agent and combining it with the first and second analysis models to parse alarm information, the problem of highly specialized alarm information in the monitoring system is solved, making alarm information easy to understand and efficiently processed, thereby improving operation and maintenance efficiency.

CN118694652BActive Publication Date: 2024-11-12CAPITAL INFORMATION TECH DEV CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411155462.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-12
Estimated Expiration
2044-08-22

AI Technical Summary

Technical Problem

When existing monitoring systems generate alarm information, the alarm information is highly technical and difficult for non-technical personnel to understand, resulting in high operation and maintenance costs and long fault response times, which affects system maintenance efficiency.

Method used

A device monitoring alarm processing method based on Zabbix is ​​adopted. Monitoring data is obtained through Zabbix-agent, alarm information is quickly parsed using the first analysis model, and then the second analysis model, which combines background information and calls the large language model, is used for in-depth analysis to output easy-to-understand analysis results.

Benefits of technology

It reduces the difficulty for operation and maintenance users to understand alarm information, improves fault response speed and operation and maintenance efficiency, and provides clear alarm information processing suggestions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118694652B_ABST
    Figure CN118694652B_ABST
Patent Text Reader

Abstract

The present application provides a device monitoring alarm processing method, device, medium and equipment based on zabbix, which belongs to the field of computer technology. The method includes: obtaining a monitoring data set of a controlled device through a zabbix-agent deployed on the controlled device; analyzing whether the controlled device has an abnormality based on the monitoring data set, and when an abnormality exists, outputting an alarm information about the controlled device; receiving an analysis instruction for the alarm information; calling a preset first analysis model to parse the alarm information, and outputting initial parsing information, which reflects the abnormal meaning represented by the alarm information; obtaining background information related to the alarm information; converting the initial parsing information and the background information into information to be analyzed that can be understood by a second analysis model; calling a second analysis model to parse the information to be analyzed, and outputting analysis result information corresponding to the alarm information. The present application improves the accuracy of understanding the alarm information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a device monitoring alarm processing method, apparatus, medium, and equipment based on Zabbix. Background Art

[0002] Amid the rapid development of computer network technology and artificial intelligence (AI), monitoring systems have become widely used as a crucial means of ensuring the secure and stable operation of various servers, network equipment, and terminal devices. This is particularly true in large, complex environments, such as within the Beijing Sub-Center area. Monitoring systems must not only cover a wide range of intelligent low-voltage devices in various business scenarios, including consumer systems, refined kitchen systems, video surveillance systems, public address systems, and access control systems, but also provide real-time, accurate monitoring data to ensure the efficient operation of these diverse intelligent low-voltage devices and systems.

[0003] Existing monitoring systems often rely on large amounts of raw data and charts to display monitoring information. However, these data are difficult for non-technical personnel to understand, and the operation interface is complex, making it difficult for non-professional users to get started.

[0004] When monitoring systems generate alarms, they often contain highly specialized terminology, sometimes even custom-defined by system developers. This requires specialized personnel to translate and interpret the alarms, converting complex monitoring data and business information into easily understandable language to facilitate effective communication between diverse user groups and professionals. However, this approach not only increases personnel costs but also prolongs response time, impacting overall system maintenance efficiency.

[0005] Based on this, it is necessary to optimize the alarm information to reduce the difficulty for relevant operation and maintenance users to understand the alarm information, improve the ability to process the alarm information, and improve operation and maintenance efficiency. Summary of the Invention

[0006] Based on this, the purpose of this application is to provide a device monitoring alarm processing method, device, medium and equipment based on Zabbix to solve at least one of the above problems.

[0007] In a first aspect, the present application provides a device monitoring alarm processing method based on Zabbix, the method comprising: obtaining a monitoring data set of the controlled device through a Zabbix-agent deployed on the controlled device; analyzing whether the controlled device has an abnormality based on the monitoring data set, and when an abnormality exists, outputting alarm information about the controlled device; receiving an analysis instruction for the alarm information; calling a preset first analysis model to parse the alarm information, and outputting initial analysis information, wherein the initial analysis information reflects the abnormal meaning represented by the alarm information; obtaining background information related to the alarm information; converting the initial analysis information and the background information into information to be analyzed that can be understood by a second analysis model; calling the second analysis model to parse the information to be analyzed, and outputting analysis result information corresponding to the alarm information, wherein the first analysis model is an analysis model that integrates knowledge base information related to the controlled device and the Zabbix framework; and the second analysis model is a large language model.

[0008] In one embodiment, the obtaining of background information related to the alarm information includes: obtaining basic device information of the controlled device; determining an abnormal time period based on the time when the alarm information is generated; extracting monitoring data generated within the abnormal time period and related to the initial parsed information from the monitoring data set, and using the basic device information and the extracted monitoring data as the background information.

[0009] In one embodiment, the acquiring of background information related to the alarm information includes: acquiring historical analysis data of historical alarms related to the alarm information, and using the historical analysis data as the background information.

[0010] In one embodiment, the obtaining of background information related to the alarm information further includes: parsing the alarm location and alarm device from the alarm information; calling a camera matching the alarm device through the RTSP protocol to capture on-site environmental information at the alarm location, and using the on-site environmental information as the background information.

[0011] In one embodiment, the second analysis model includes multiple models. Before converting the initial parsing information and the background information into information to be analyzed that can be understood by the second analysis model, the method further includes: determining at least one second analysis model to be used based on the initial parsing information and the background information.

[0012] In one embodiment, the converting of the initial parsed information and the background information into information to be analyzed that can be understood by the second analysis model includes: disassembling multiple sub-data from the initial parsed information and the background information, each sub-data matching one or more second analysis models; and converting each sub-data into sub-information to be analyzed that can be understood by the matching second analysis model based on the analysis template of the second analysis model.

[0013] In one embodiment, the calling of the second analysis model to parse the information to be analyzed and outputting analysis result information corresponding to the alarm information includes: determining the parsing order of each sub-data based on the multiple sub-data; calling the second analysis model corresponding to each sub-data to process the sub-information to be analyzed according to the parsing order; and using the analysis result sub-information of the preceding sub-data among the multiple sub-data with a dependent relationship as the input of the second analysis model corresponding to the succeeding sub-data; and fusing the analysis result sub-information output by each second analysis model to form the analysis result information.

[0014] In one embodiment, the controlled device includes an access control device; the method further includes: when it is determined based on the analysis result information that the access control device needs to be remotely controlled, obtaining image data captured by a camera associated with the access control device; identifying the image data, and remotely controlling the opening and release of the access control device based on the identification result.

[0015] In a second aspect of the present application, a device monitoring and alarm processing device based on Zabbix is ​​provided, the device comprising:

[0016] The monitoring data acquisition module is used to obtain the monitoring data set of the controlled device through the zabbix-agent deployed on the controlled device;

[0017] an alarm identification module, configured to analyze whether the controlled device has an abnormality based on the monitoring data set, and output alarm information about the controlled device when an abnormality exists;

[0018] An analysis instruction acquisition module, configured to receive an analysis instruction for the alarm information;

[0019] An initial parsing module, configured to call a preset first analysis model to parse the alarm information and output initial parsing information, wherein the initial parsing information reflects the abnormal meaning indicated by the alarm information;

[0020] A background information acquisition module, used to obtain background information related to the alarm information;

[0021] A secondary analysis module is used to convert the initial analysis information and the background information into information to be analyzed that can be understood by a second analysis model; call the second analysis model to parse the information to be analyzed, and output analysis result information corresponding to the alarm information. The first analysis model is an analysis model that integrates knowledge base information related to the controlled device and the zabbix framework; the second analysis model is a large language model.

[0022] In a third aspect of the present application, a computer-readable storage medium is provided, on which executable instructions are stored. When the executable instructions are executed by a processor, the processor executes the method described in any embodiment of the present application.

[0023] In a fourth aspect of the present application, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more programs, which, when executed by the one or more processors, enables the one or more processors to execute the method described in any embodiment of the present application.

[0024] The device monitoring alarm processing method, device, medium and equipment based on Zabbix in this application monitor the controlled devices by adopting the Zabbix framework, so that a large number of controlled devices can be monitored conveniently and quickly, and the monitoring information of multiple dimensions such as the current device status and historical operating status of the controlled devices can be presented in a visual form, so that relevant staff can clearly and intuitively understand the current operating status and historical operating status of the controlled devices.

[0025] Furthermore, when an abnormality is detected in the controlled device, an alarm message can be generated in time, and the alarm message can be analyzed by setting two analysis models. Specifically, the first analysis model that integrates the relevant knowledge base of the zabbix framework is first used to quickly parse the alarm message, thereby obtaining initial analysis information, and then after obtaining background technology related to the alarm information, the second analysis model is called to analyze the background technology and the initial analysis information, so that the second analysis model can obtain a more accurate or comprehensive analysis result relative to the initial analysis information based on the initial analysis information through the rich background information related to the alarm information provided, so that relevant users can easily and conveniently understand the meaning of the alarm information and the response suggestions, reducing the difficulty of relevant operation and maintenance users in understanding the alarm information, and providing a strong technical guarantee for the user's subsequent alarm processing and decision-making on the monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope of the present application.

[0027] Figure 1 The following is a schematic diagram of a scenario of a device monitoring alarm processing method based on Zabbix in one embodiment;

[0028] Figure 2 The following is a flow chart of a device monitoring alarm processing method based on Zabbix in one embodiment;

[0029] Figure 3 A schematic diagram of a process for obtaining background information related to alarm information in one embodiment;

[0030] Figure 4 A schematic diagram of a process for calling a second analysis model to parse information to be analyzed and outputting analysis result information corresponding to alarm information in one embodiment;

[0031] Figure 5 This is a schematic diagram of the structure of a device monitoring and alarm processing device based on Zabbix in one embodiment;

[0032] Figure 6 This is a schematic diagram of the structure of a device monitoring and alarm processing device based on Zabbix in another embodiment;

[0033] Figure 7 FIG. 4 is a structural block diagram of an electronic device in an embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0035] All terms (including technical and scientific terms) used in this application have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0036] For example, the terms "first," "second," and the like may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, a first analysis model may be referred to as a second analysis model, and similarly, a second analysis model may be referred to as a first analysis model without departing from the scope of this application. The first analysis model and the second analysis model are both analysis models, but they are not the same analysis model.

[0037] For example, the terms "include", "comprising", etc. used in this application indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0038] The device monitoring alarm processing method based on Zabbix in this application can be applied to Figure 1 In the scenario shown. Combined Figure 1 As shown, the electronic device 110 can be a server deployed with zabbix-server, and can remotely monitor multiple controlled devices 120, and form visual information based on the monitoring data. Among them, when the number of controlled devices 120 is large, the electronic device 110 can reduce the pressure of zabbix-server by installing the agent zabbix-proxy as needed. By using the distributed monitoring environment through zabbix-proxy, the agent collects the remote location data of the controlled device 120 and sends the data to the server. This can reduce the load on the server and improve monitoring efficiency. The electronic device 110 is also provided with a storage space for storing data. The monitored data and the configuration information of all controlled devices can be stored in a set database, such as a MySql database.

[0039] The monitoring system installed on the electronic device 110, based on the Zabbix framework configuration, can implement hardware monitoring, system operation monitoring, software operation status parameters, database status, log monitoring, automatic recovery functions, alarm push, and other functions for multiple controlled devices 120. Among them, a corresponding Zabbix-agent is deployed on each controlled device 120, and the electronic device 110 monitors various data through the Zabbix-agent.

[0040] The electronic device 110 analyzes the acquired monitoring data and presents it visually. For example, it can generate corresponding charts based on various monitoring data and present them as icons on a web page. Furthermore, it can display information via the web and transmit the received monitoring data, collected video information, or generated visualization information to the video monitoring client 130 via relevant transmission protocols for information presentation. The electronic device 110 also receives processing instructions sent by the video monitoring client 130 and processes the electronic device 110 and the controlled device 120 based on the processing instructions.

[0041] Furthermore, the electronic device 110 is pre-configured with a corresponding robot communication platform (e.g. Figure 1 The Langchain-Chatchat shown in the figure is used to integrate the corresponding analysis model. The analysis task is output to the analysis model through the robot communication platform, so that the analysis model can parse the analysis task and obtain the corresponding analysis result information.

[0042] For example, the electronic device detects whether there is an anomaly in the controlled data and generates an alarm message when an anomaly is identified. Based on the generated alarm information, the electronic device can call the corresponding first analysis model and / or second analysis model through a pre-built robot communication platform to perform the analysis task of parsing the alarm information, and obtain the analysis result information output by the relevant analysis model, and display the analysis result information on the web, or send it to the video surveillance client, so that relevant personnel can clearly understand the meaning of the alarm information and how to handle the alarm information.

[0043] In one embodiment, Figure 2 As shown, a device monitoring alarm processing method based on Zabbix is ​​provided. The method includes:

[0044] Step 202: Obtain the monitoring data set of the controlled device through the zabbix-agent deployed on the controlled device.

[0045] In this embodiment, there may be multiple controlled devices, each of which is deployed with a zabbix-agent. Electronic devices can use this zabbix-agent to obtain relevant monitoring data from the controlled devices. Specifically, controlled devices can be devices in automated control and monitoring systems that can be controlled, monitored, and managed by remote or automated control systems. These devices typically have the ability to communicate with the control system to receive control instructions, send status information, or perform preset tasks. Controlled devices are widely used in various industrial, commercial, and residential environments to improve efficiency, safety, and convenience. For example, these devices can be network devices, servers, industrial computers, and other equipment. They can also be one or more related access control devices such as access controllers, visitor terminals, vehicle recognition cameras, facial recognition panels, facial recognition columns, personnel verification machines, LED displays, gate controllers, verification machines, and entrance and exit cameras. Electronic devices can monitor all relevant controlled devices in a region using the zabbix framework. For example, multiple weak current intelligent devices in a region such as the Beijing Urban Sub-Center can be monitored.

[0046] The monitoring data set contains multiple pieces of monitoring data. Monitoring data can be divided into hardware data, system operation-related data, software operation-related status parameters, database status data, log data, and other monitoring types based on monitoring type. For example, monitoring data can include one or more of network connectivity, operating system CPU usage and load indicators, physical memory and virtual memory usage, disk space usage and read and write speeds, network interface inbound and outbound traffic, errors and packet loss, service status, application, security, and system logs. It can also include one or more of specific process status, server hardware status (such as temperature, fan speed, etc.), database performance and status, operating system version, host name, system uptime, and one or more of network device connection traffic and status, interface type and rate, routing status and changes, network latency and packet loss.

[0047] Relevant monitoring data is configured on the front end and stored in a database. For example, monitoring data can be collected from the monitoring server using Zabbix-agent and stored in a MySQL database. The collected monitoring data can include the corresponding data generation or collection time. This allows data collected for a period of time prior to the current time to be used as historical data. Trend data can be generated based on the data collected at each time point. The collected historical and trend data are also stored in the database.

[0048] Historical data typically includes recent monitoring data, while trend data covers monitoring data over longer timeframes. This data can be used to generate graphs and reports, as well as for historical analysis. Event and alarm data generated from monitoring data, such as trigger status changes and notification history, is also stored in the database. This data is invaluable for troubleshooting and performance analysis.

[0049] Electronic devices create corresponding monitoring items for each controlled device through Zabbix-Server and obtain corresponding monitoring data based on these monitoring items. Specifically, based on the status of the controlled device and the type of monitoring data to be collected, the electronic device can use active or passive inspection methods to obtain the corresponding controlled device's monitoring data, forming the corresponding monitoring data set.

[0050] Active checks involve the Zabbix server or agent periodically sending monitoring requests to a controlled device configured with the Zabbix-agent, obtaining relevant monitoring data and forming a monitoring data set. Passive checks involve the controlled device proactively sending monitoring data to the Zabbix server or agent via the Zabbix-agent deployed on it. This method is suitable for controlled devices that cannot be directly accessed from the Zabbix server or for monitoring data that is inconvenient to actively monitor. For example, log monitoring data can be obtained through passive checks.

[0051] Step 204: Analyze whether the controlled device has any abnormality based on the monitoring data set, and output alarm information about the controlled device when an abnormality exists.

[0052] In this embodiment, the electronic device further sets corresponding triggers for each controlled device through Zabbix-Server. When the value of the monitoring item (i.e., the corresponding monitoring data) exceeds a preset threshold, the trigger will be activated and an alarm will be generated.

[0053] For the acquired monitoring data set, the electronic device can classify and store it, and analyze the stored relevant monitoring data to identify whether the controlled device corresponding to the monitoring data has an abnormality. Specifically, the electronic device pre-sets the alarm conditions for each controlled device, and compares and analyzes one or more monitoring data monitored by the controlled device with the set alarm conditions. When it is identified that the corresponding alarm condition is met, the trigger will be activated, and it will be determined that the corresponding controlled device has an abnormality and generate an alarm message. For example, when the corresponding value of one or more monitoring indicators of a device exceeds the set safety value, it is determined that the corresponding controlled device has an abnormality.

[0054] The alarm information includes the device identifier of the abnormal controlled device and the relevant alarm content. This alarm content is typically system-defined or contains a relevant alarm code. Alarm content can include device failures, network anomalies, performance bottlenecks, security incidents, and other information, guiding operations and maintenance personnel in locating the problem. For example, the alarm content might be "Unavailable by ICMP ping" or a predefined alarm code indicating "Unavailable by ICMP ping."

[0055] Furthermore, in addition to the device identifier and alarm content, the alarm information may also include one or more relevant information, including the alarm level, alarm location information, and alarm time. Alarm levels may include levels such as severe, general, and minor, reflecting the urgency and severity of the issue. In a monitoring system, if an alarm is associated with a specific geographic or logical location, this location information may be included in the alarm information to help operations personnel understand the specific circumstances surrounding the alarm. The alarm time typically includes a date and timestamp, which helps operations personnel track and analyze the occurrence and development of the problem.

[0056] Step 206: Receive an analysis instruction for the alarm information.

[0057] In this embodiment, the electronic device can display the generated alarm information on a web-based visual display, allowing relevant users to view the corresponding alarm information on the web display page. In addition, the web display interface also provides corresponding touch buttons for analyzing the meaning of the alarm information and the related causes. Users can perform touch operations on these touch buttons. For example, relevant touch buttons such as "Cause Analysis" and "Solution" for the alarm information can be provided on the web display interface or on the video surveillance client. When a touch operation on the corresponding touch button is detected, an analysis instruction for the alarm information can be triggered.

[0058] Furthermore, the electronic device can send it to the corresponding client, and the client personnel can decide whether to further analyze the alarm information. Figure 1 The video surveillance client 130 in.

[0059] Optionally, the client connected to the electronic device may include one or more clients, and different clients receive corresponding information about different controlled devices. A correspondence between different client identifiers and clients can be pre-established. After detecting the device identifier of the controlled device in the alarm information, the client identifier of the corresponding client can be queried based on the device identifier, and the alarm information can be sent to the corresponding client, so that the client user can be immediately informed of the abnormal situation of the corresponding controlled device.

[0060] Specifically, electronic devices can assign corresponding clients to controlled devices based on their location and device type. For example, a correspondence can be established between controlled devices in the same or similar locations, or controlled devices of the same type, and the client identifier of a client, thereby improving the standardization of the management of controlled devices.

[0061] Step 208: Call a preset first analysis model to analyze the alarm information and output initial analysis information.

[0062] In this embodiment, the first analysis model is an analysis model that integrates the knowledge base information related to the controlled device and the zabbix framework. The initial parsing information reflects the abnormal meaning represented by the alarm information. Furthermore, the initial parsing information can also include the preliminary determination of the cause of the abnormal situation and the preliminary solution to the abnormal problem. The knowledge base collects the meanings represented by various alarm information and preliminary solutions. For the generated alarm information, the electronic device can query the meaning represented by the alarm information from the preset first analysis model according to the analysis instruction. For example, the initial parsing information parsed for the alarm information is the meaning of the relevant abnormality such as "controlled device network abnormality", "controlled device disk space abnormality", "device cannot be pinged through ICMP protocol", etc.

[0063] For example, for the above-mentioned alarm information "Unavailable by ICMP ping", the initial parsing information obtained by the first analysis model may include "Unavailable by ICMP ping means that the device cannot be pinged through the ICMP protocol. Solution: check the network connection, firewall settings, device status or restart the device."

[0064] Specifically, the electronic device can call the first analysis model through the above-mentioned Langchain-Chatchat tool, use the warning information as the input of the first analysis model, and receive the initial analysis information parsed by the first analysis model. Figure 1As shown, the first analysis model is an independent service running in the intranet. After receiving the alarm information through the corresponding application programming interface (API), the Langchain-Chatchat tool communicates with the service through the intranet IP address and port number, thereby transmitting the alarm information to the first analysis model. It also receives the initial parsing information generated by the first analysis model based on the analysis of the alarm information.

[0065] By using the first analysis model to perform initial analysis of the alarm information, the abnormal meaning represented by the alarm information can be quickly obtained.

[0066] Step 210: Obtain background information related to the alarm information.

[0067] In this embodiment, after obtaining the alarm information, the electronic device may further obtain background information related to the alarm information, so that the alarm information can be further analyzed based on the background information to identify the cause of the alarm information or a solution.

[0068] Context information refers to information associated with the alarm information and / or helpful for understanding or processing the alarm information. Context information includes basic device information of the controlled device, monitored log files, and related monitoring data. Obtaining relevant context information provides more comprehensive information for subsequent analysis of the alarm information. Basic device information includes one or more of the following: device name, device type, device model, and device configuration information. Related monitoring data can be obtained from the acquired monitoring data set, specifically monitoring data related to the alarm information.

[0069] The data types of background information may include various types, such as text, image, audio, and video. For example, text data may be the aforementioned monitoring data of physical memory and virtual memory; image data may be captured monitoring images related to the controlled device, such as images containing suspected fault locations or images of the operating environment; audio data may be sound data of the monitored controlled device during operation; and video data may be video data of the monitored controlled device during operation.

[0070] In one embodiment, background information related to the alarm information is obtained based on the initial parsed information.

[0071] For example, the initial parsing information obtained based on the above-mentioned alarm information "Unavailable by ICMP ping" may be related to the network, firewall, and device status. The electronic device can then further obtain information such as network parameters, firewall parameters, and device status parameters related to the controlled device and use this information as background information.

[0072] Step 212: Convert the initial parsed information and background information into information to be analyzed that can be understood by the second analysis model.

[0073] In this embodiment, the second analysis model is a large language model. For example, the large language model may be one or more of a general large model, a domain-specific large model, and a multimodal large model. For example, the general large model may include GPT series models, BERT models, Qwen, chatglm, etc., and the domain-specific large model may include the Transformer model in the field of computer vision and the XLNet model in the field of natural language processing. The multimodal large model includes the CLIP model and the DALL-E model, etc. Compared with the first analysis model, the second analysis model can provide more comprehensive analysis results. And the analysis results obtained by the second analysis model are very dependent on the specific information provided and the questioning method for the second analysis model. Generally speaking, although the same information is provided, if different questioning methods are used, the analysis results obtained may not be the same.

[0074] This application integrates the obtained initial parsed information and background information into information to be analyzed that can be understood by the second analysis model. The information to be analyzed includes information that needs to be analyzed by the second analysis model and corresponding questioning methods.

[0075] In one embodiment, there are multiple second analysis models, each of which contains one or more model attribute information, including the data type it excels at processing, the technical field it excels at processing, and the associated resources it consumes. Based on the model attribute information of the second analysis models, one or more matching second analysis models can be selected from the background information and initial parsed information, thereby generating information to be analyzed that is suitable for processing according to the selected second analysis models.

[0076] Specifically, the electronic device is pre-configured with multiple analysis templates, each applicable to one or more data types and one or more types of data within that data type. For example, the template includes a question format specific to the second analysis model, so that the corresponding second analysis model generates optimal analysis results when processing the information to be analyzed generated based on the question format.

[0077] Specifically, the analysis template defines the data filling specification information of the sub-data and the specification information of the analysis instructions for the second analysis model. The data filling specification information includes how the data content of different sub-data is integrated into the analysis template, and the specification information of the analysis instructions includes one or more of the standard questioning method for the second analysis model and the analysis constraints and analysis results that the second analysis model needs to provide. The analysis constraints may include what rules need to be followed for analysis, such as an agreement that an analysis result of less than 100 words needs to be generated; the analysis result specifies what content needs to be output and in what form the content needs to be output, such as indicating that the analysis result needs to be output in text form or image form.

[0078] The initial parsed information and background information can be parsed to identify the data type and data content, and then the most suitable analysis template is selected based on the data type and data content, and then the initial parsed information and background information are converted into information to be analyzed based on the analysis template.

[0079] In one embodiment, the analysis template may be a template optimized through training and testing. The electronic device pre-prepares multiple initial templates and sample data sets, where each initial template corresponds to one or more data types, and each sample data belongs to one of the data types. For sample data and initial templates of the same data type, the relevant training model is invoked to iteratively train the initial template, outputting an optimized initial template. The optimized initial template is then used as the analysis template for the corresponding data type.

[0080] Specifically, the training model can be a related convolutional neural network model. For the template to be iterated, the corresponding data sample is filled into the iterative template to form the corresponding sample information to be analyzed, and the sample information to be analyzed is input into the corresponding second analysis model for operation, and the corresponding sample analysis result is output, and the sample information result is compared with the pre-prepared standard result, in which each sample data corresponds to a standard result. The difference between the sample information result and the standard result is detected, and the template to be iterated is adjusted according to the difference, so as to continue to iterate according to the above process according to the adjusted template to be iterated, until the number of iterations reaches the preset number threshold, or the difference between the obtained sample information result and the standard result is less than the preset difference threshold. Wherein, in the first iteration, the template to be iterated is the set initial template.

[0081] Step 214: Call the second analysis model to parse the information to be analyzed, and output analysis result information corresponding to the alarm information.

[0082] In this embodiment, the second analysis model is an external model. The electronic device can interact with the corresponding online server of the second analysis model through HTTP requests, pass the information to be analyzed to the second analysis model, and instruct the second analysis model to process the information to be analyzed, generate corresponding analysis results, and transmit the analysis results to the electronic device through a preconfigured interface. The electronic device can then obtain the analysis results output by the second analysis model.

[0083] Taking the alarm "Unavailable by ICMP ping" as an example, by combining the initial parsed information with the contextual information generated to generate the analysis information, the second analysis model can generate a more accurate analysis of the cause of the problem and a solution. For example, the cause might be "most likely a fault in the XX application," and the solution might be "please restart the XX application."

[0084] It is understandable that for different controlled devices, even if the alarm content generated is the same, the corresponding solution and other analysis results may not be the same. However, the present application uses two analysis models. First, the abnormal meaning represented by the alarm information is obtained with the help of the first analysis model, and then the abnormal meaning analyzed and the background information related to the alarm information are combined into information to be analyzed that can be understood by the second analysis model. The information to be analyzed provided to the second analysis model is not only the alarm information, but also more background information related to the alarm information. Therefore, the second analysis model can parse out more accurate and / or comprehensive analysis results based on richer information, so that relevant users can easily and conveniently understand the meaning and response suggestions represented by the alarm information, reduce the difficulty of relevant operation and maintenance users in understanding the alarm information, improve the ability to process the alarm information, and improve operation and maintenance efficiency.

[0085] In one embodiment, Figure 3 As shown, step 210 includes:

[0086] Step 302: Obtain basic device information of the controlled device.

[0087] In this embodiment, based on the generated alarm information, the system can determine which controlled device has issued the alarm and can then obtain basic device information for the controlled device. Basic device information includes one or more of the following: basic device information, technical parameter information, and device maintenance information. Basic device information includes one or more of the following: device name, device code, and device specifications and models. The device specifications and models reflect the device's dimensions, performance, and other characteristics. Technical parameter information includes one or more of the following: the device's operating conditions, performance parameters, and other technical details. Device maintenance information includes information such as the time, content, and results of previous maintenance operations.

[0088] Step 304: Obtain the abnormal time period based on the generation time of the alarm information.

[0089] In this embodiment, the start time of the abnormal period can be the time when the alarm information is generated, or the time when the monitoring indicator associated with the alarm information first exceeds the corresponding safe value range within a period of time before the generation time. The time period from the start time to the current time is regarded as the abnormal period.

[0090] Furthermore, the generation time corresponding to the alarm information can be obtained, and the corresponding analysis duration can be determined based on the initial parsing information. The time corresponding to the analysis duration before the generation time can be determined, and this time can be used as the starting time of the abnormal period. For example, if the alarm information is generated at 10:10:15, and the analysis duration required to be considered is determined to be 2 hours based on the initial parsing information, the corresponding starting time can be determined to be 8:10:15. The analysis durations determined by different initial parsing information may not be the same.

[0091] Optionally, there is no limitation on the execution order between step 302 and step 304. For example, they may be executed in parallel, or step 302 may be executed first and then step 304.

[0092] In one embodiment, the abnormal period is determined based on the basic device information and the alarm information, and the abnormal period is obtained.

[0093] Optionally, the abnormal period is also related to the basic information of the controlled device and the device. For example, for the same alarm information, the abnormal period required to be obtained for different device types may not be the same. By combining the alarm information and basic device information to determine the abnormal period, the obtained background information can be more closely matched to the controlled device, thereby improving the accuracy and comprehensiveness of subsequent secondary analysis of the alarm information.

[0094] Step 306 : extracting monitoring data generated during the abnormal period and related to the initial parsed information from the monitoring data set, and using the basic device information and the extracted monitoring data as background information.

[0095] When the abnormal period is determined, monitoring data related to the initial parsed information monitored for the controlled device during the abnormal period can be extracted, and the extracted monitoring data and basic information can be used as background information.

[0096] Among them, the initial parsing information includes the meaning represented by the alarm information. For example, when it indicates that there is a network failure, the relevant monitoring data may be data on monitoring parameters related to the network failure. Furthermore, the monitoring data related to the initial parsing information also includes data corresponding to abnormal monitoring indicators.

[0097] For example, for the initial parsing information corresponding to the above-mentioned alarm information "Unavailable by ICMP ping", the extracted monitoring data may involve data on parameters related to the ICMP protocol, firewall status, network connection status, device status, etc.

[0098] By using basic equipment information and monitoring parameters related to the initial analysis information during the abnormal period as background technology, the second analysis model can parse the alarm information based on more information, thereby obtaining a more accurate cause of the problem and the corresponding treatment method.

[0099] In one embodiment, step 210 further includes: obtaining historical analysis data of historical alarms related to the alarm information, and using the historical analysis data as background information.

[0100] In this embodiment, historical alarms include alarms identical to the alarm that occurred within a historical time period, as well as alarms generated by the controlled device within that time period. The historical time period can be any appropriate time period, such as the last year or the last three months. It can also be a time period adaptively set based on initial analysis information. It is understood that different faults may not necessarily require the same historical time period for consideration.

[0101] Historical analysis data represents the analysis results corresponding to the historical alarm information. This analysis result data is accurate data after verification. For example, if there is a historical alarm information generated, the historical analysis data includes the basic device information of the historical alarm device, as well as the alarm analysis results and processing results.

[0102] By further combining historical analysis data, the second analysis model can use more information to analyze the alarm information, and the obtained analysis results are more accurate.

[0103] In one embodiment, step 210 further includes: parsing the alarm location and alarm device from the alarm information; and capturing on-site environment information at the alarm location by calling a camera that matches the alarm device via the RTSP protocol, and using the on-site environment information as background information.

[0104] Not all alarms require the use of on-site environmental information. The electronic device pre-sets alarm information that requires obtaining on-site environmental information. After step 204, it determines whether to capture on-site environmental information based on the alarm information. When obtaining on-site environmental information is required, this embodiment is executed.

[0105] For example, if the alarm indicates a suspected device intrusion, physical damage, or tampering with device-related data, or a critical service interruption or large-scale failure, then it is determined that on-site environmental information needs to be obtained. Alternatively, if the controlled device corresponding to the alarm is identified as being captured by a camera, the on-site environmental information of the controlled device can also be obtained, thereby obtaining more comprehensive and accurate contextual information related to the alarm.

[0106] Specifically, the alarm location (i.e., the specific location or area where the alarm occurs) and the alarm device are parsed from the received alarm information. The alarm device can be the controlled device itself or a related component deployed on the controlled device, such as a smoke detector, temperature sensor, etc. deployed on the related controlled device.

[0107] For the determined controlled device, the camera related to the controlled device can be queried from the internal database or resource management system, and the camera corresponding to the device of this type can be matched.

[0108] After matching a suitable camera, the system uses the Real-Time Streaming Protocol (RTSP) to send a request to the camera, instructing it to capture and transmit images and videos containing the on-site environment at the alarm location, and use the images or videos as the on-site environment information and as part of the background technology.

[0109] In one embodiment, the second analysis model includes multiple models. Before step 212, the method further includes: determining at least one second analysis model to be used based on the initial parsed information and the context information. Step 212 includes: extracting multiple sub-data from the initial parsed information and the context information, each sub-data being matched to one or more of the second analysis models; and converting each sub-data into sub-information to be analyzed that is understandable by the matched second analysis model based on an analysis template of the second analysis model.

[0110] In this embodiment, since there are multiple data types in the initial parsed information and / or background information, and the second analysis model is difficult to accurately process all data types, multiple second analysis models can be set to determine different second analysis models corresponding to different data types.

[0111] Specifically, since the initial parsed information and / or background information contain a large amount of data, and there may be data corresponding to different data types, such as text data, image data, and audio data. The electronic device can decompose the data in the initial parsed information and the background information to form multiple sub-data. Among them, there may be some identical data in the sub-data, for example, the decomposed sub-data contain all or part of the basic device information of the controlled device. Each sub-data belongs to a data type and can be processed independently by a corresponding second analysis model. The sub-task only contains a task that requires the second analysis model to analyze and data related to the task, and the sub-task cannot be further decomposed.

[0112] The electronic device breaks down the data into multiple sub-data based on the analysis instruction, the initial parsing information and the background information.

[0113] For each piece of decomposed sub-data, the degree of matching between the sub-data and multiple second analysis models can be calculated based on the data content of the sub-data and the task to be parsed, combined with the model attribute information of each of the aforementioned models. The second analysis model with the highest matching degree or exceeding a certain matching degree threshold is selected as the analysis model that matches the sub-data. The matching degree threshold can be a pre-set value of any appropriate size.

[0114] In one embodiment, a related cosine similarity calculation model may be set, and the cosine similarity between the two may be calculated based on the model attribute information and the data content of the sub-data, and the matching degree between each sub-data and the second analysis model may be determined based on the cosine similarity.

[0115] For the determined matching second analysis model, the analysis template corresponding to the second analysis model can be called. According to the analysis template, the sub-data is converted into the corresponding sub-information to be analyzed, so that the second analysis model can output a more accurate analysis result for the sub-information to be analyzed.

[0116] In one embodiment, Figure 4 As shown, step 214 includes:

[0117] Step 402: Determine the parsing order of each sub-data based on the multiple sub-data.

[0118] In this embodiment, since multiple sub-data are decomposed from the initial parsing information and background information, and are ultimately used for the analysis of the alarm information, there is a certain degree of correlation among these sub-data, and there is also dependency between some of the sub-data. Among them, the electronic device can determine the corresponding dependency relationship based on the content between the various sub-data, and determine the parsing order of each sub-data based on the dependency relationship. For example, when the analysis and processing of one of the two sub-data requires the use of the analysis result of the other sub-data, it means that there is a dependency relationship between the two sub-data. When a sub-data needs to rely on the parsing results of one or more other sub-data, the parsing order of the sub-data is sorted after the parsing of the one or more sub-data on which it depends. For sub-data that do not have a dependency relationship, the parsing order can be unrestricted. For example, it can be executed in parallel based on the processing power of the current system, or the electronic device can sort the unrestricted sub-data in an appropriate parsing order based on the resource usage of the current system.

[0119] Step 404: Call the second analysis model corresponding to each sub-data according to the parsing order to process the sub-information to be analyzed.

[0120] As described above, the electronic device determines the second analysis model used to process the corresponding sub-data based on the calculated matching degree between each sub-data and the second analysis model. Specifically, for a certain sub-data, when there is only one second analysis model whose matching degree reaches the matching degree threshold, it is directly determined to call the second analysis model to process the sub-data; when there are multiple second analysis models whose matching degree reaches the matching degree threshold, the electronic device can call the second analysis model with the highest response efficiency from the multiple second analysis models to process the sub-data. In the case where there are multiple second analysis models that meet the calling conditions, the electronic device can calculate which second analysis model to use to process the corresponding sub-data based on the current situation of all sub-data, so that the overall analysis result information can be obtained as quickly as possible (that is, the overall response efficiency), and based on this, it is determined which second analysis model to call to process the corresponding sub-data. Optionally, a greedy algorithm can be used to calculate the corresponding overall response efficiency.

[0121] For example, at the current moment, there are 10 sub-data that can be processed simultaneously, among which the first 5 sub-data are determined to be only able to be processed using the second analysis model A, and for the 6th sub-data, the second analysis model that meets the matching threshold includes the second analysis model A and the second analysis model B. At this time, if the second analysis model A is used for processing, then at this moment, the second analysis model needs to process 6 sub-data. Since the response speed is lower than the second analysis model B due to the excessive number of sub-data that need to be processed, the second analysis model B can be called to process the 6th sub-data.

[0122] Step 406: Use the analysis result sub-information of the preceding sub-data among the plurality of sub-data with a dependency relationship as the input of the second analysis model corresponding to the succeeding sub-data.

[0123] In this embodiment, the analysis result sub-information is the output information after the second analysis model processes the corresponding sub-data. The electronic device sets a corresponding result storage container for the alarm information, and the result storage container is used to store the analysis result sub-information corresponding to each sub-data. For the sub-data with a later parsing order that has a dependency relationship, the electronic device can obtain the processing result sub-information of all the sub-data it depends on from the result storage container based on the dependency relationship, and pass the sub-data and the processing result sub-information to the corresponding second analysis model through the communication interface set with the second analysis model, and receive the processing result sub-information of the sub-data returned by the second analysis model through the interface, and continue to store the processing result sub-information in the result storage container.

[0124] Step 408: Perform fusion processing on the analysis result sub-information output by each second analysis model to form analysis result information.

[0125] In this embodiment, after processing all sub-data, the electronic device can fuse the processed sub-information of each sub-data, so that the final analysis result information corresponds to the analysis instruction. The fusion process may include operations such as integrating, deduplicating, sorting, and deleting the individual analysis result sub-information to ensure that the final analysis result information is consistent and readable and meets the requirements of the analysis instruction. Through fusion processing, a comprehensive and accurate analysis result can be obtained, providing a strong technical support for the user's subsequent alarm processing and decision-making regarding the monitoring system.

[0126] In one embodiment, the controlled device includes an access control device; the method further includes: when it is determined based on analysis result information that the access control device needs to be remotely controlled, obtaining image data captured by a camera associated with the access control device; identifying the image data, and remotely controlling the opening and release of the access control device based on the identification result.

[0127] In this embodiment, the access control device refers to a device that can control whether to open and release by capturing the recognition results with a camera. It can include intelligent image recognition access control devices, face recognition access devices, license plate recognition release devices, visual case access devices, visual recognition gate machines and other controlled devices.

[0128] Specifically, when the analysis result information indicates that the access control device meets the remote access control conditions, it can communicate with the camera related to the access control device through the relevant protocol, receive the real-time image data uploaded by the camera, and combine the relevant information stored in the database to identify and detect the received image data to obtain the judgment result of whether to open or close the access control device, and control the opening and release of the access control device based on the judgment result.

[0129] Specifically, when there are multiple access control devices in the system that require remote access control, due to the constraints of the system's available resources, the electronic device can prioritize each access control device. The priority ranking can be based on the severity of the problem reflected in the alarm information, and for each different type of access control device, different image or video compression algorithms are used for transmission according to the corresponding real-time data transmission requirements, thereby achieving optimal utilization of computing resources.

[0130] In one embodiment, Figure 5 As shown, a device monitoring and alarm processing device based on Zabbix is ​​provided, which includes:

[0131] The monitoring data acquisition module 502 is used to obtain the monitoring data set of the controlled device through the zabbix-agent deployed on the controlled device;

[0132] An alarm identification module 504 is used to analyze whether a controlled device has an abnormality based on the monitoring data set, and output alarm information about the controlled device when an abnormality exists;

[0133] An analysis instruction acquisition module 506 is used to receive analysis instructions for alarm information;

[0134] The initial parsing module 508 is configured to call a preset first analysis model to parse the alarm information and output initial parsing information, which reflects the abnormal meaning indicated by the alarm information;

[0135] A background information acquisition module 510 is used to obtain background information related to the alarm information;

[0136] The secondary analysis module 512 is used to convert the initial analysis information and background information into information to be analyzed that can be understood by the second analysis model; call the second analysis model to analyze the information to be analyzed, and output analysis result information corresponding to the alarm information.

[0137] Among them, the first analysis model is an analysis model that integrates the knowledge base information related to the controlled device and the zabbix framework; the second analysis model is a large language model.

[0138] In one embodiment, the background information acquisition module 510 is also used to obtain basic device information of the controlled device; obtain the abnormal time period based on the generation time of the alarm information; extract the monitoring data generated during the abnormal time period and related to the initial parsed information from the monitoring data set, and use the basic device information and the extracted monitoring data as background information.

[0139] In one embodiment, the background information acquisition module 510 is further configured to acquire historical analysis data of historical alarms related to the alarm information, and use the historical analysis data as background information.

[0140] In one embodiment, the background information acquisition module 510 is also used to parse the alarm location and alarm device from the alarm information; through the RTSP protocol, the camera matching the alarm device is called to capture the on-site environment information at the alarm location, and the on-site environment information is used as background information.

[0141] In one embodiment, the second analysis model includes a plurality.

[0142] In one embodiment, the secondary analysis module 512 is further configured to determine at least one second analysis model to be used based on the initial analysis information and the background information.

[0143] In one embodiment, the secondary analysis module 512 is also used to disassemble multiple sub-data from the initial analysis information and background information, each sub-data matches one or more second analysis models; and based on the analysis template of the second analysis model, each sub-data is converted into sub-information to be analyzed that can be understood by the matching second analysis model.

[0144] In one embodiment, the secondary analysis module 512 is also used to determine the analysis order of each sub-data based on multiple sub-data; call the second analysis model corresponding to each sub-data to process the sub-information to be analyzed according to the analysis order; and use the analysis result sub-information of the previous sub-data among the multiple sub-data with a dependent relationship as the input of the second analysis model corresponding to the subsequent sub-data; and fuse the analysis result sub-information output by each second analysis model to form analysis result information.

[0145] In one embodiment, the controlled device includes an access control device. Figure 6 As shown, the device also includes:

[0146] The release control module 514 is used to obtain image data captured by a camera related to the access control device when it is necessary to remotely control the access control device based on the analysis result information; identify the image data, and remotely control the opening and release of the access control device based on the identification result.

[0147] In one embodiment, a computer-readable storage medium is provided, on which executable instructions are stored. When the instructions are executed by a processor, the processor executes the steps in the above-mentioned method embodiments.

[0148] In one embodiment, an electronic device is also provided, comprising one or more processors; a memory, wherein one or more programs are stored in the memory, wherein when the one or more programs are executed by one or more processors, the one or more processors execute the steps in the above-mentioned method embodiments.

[0149] In one embodiment, Figure 7 , which shows a schematic diagram of the structure of an electronic device for implementing an embodiment of the present application. Electronic device 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 702 or programs loaded from storage unit 708 into random access memory (RAM) 703. RAM 703 also stores various programs and data required for the operation of electronic device 700. CPU 701, ROM 702, and RAM 703 are connected to each other via bus 704. Input / output (I / O) interface 705 is also connected to bus 704.

[0150] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, mouse, and the like; an output section 707 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 708 including devices such as a hard disk; and a communication section 709 including a network interface card such as a LAN card or a modem. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. Removable media 711, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 710 as needed, so that computer programs read from the media can be installed in the storage section 708 as needed.

[0151] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer-readable medium carrying instructions. In such embodiments, the instructions can be downloaded and installed from a network via communication component 709 and / or installed from removable media 711. When the instructions are executed by central processing unit (CPU) 701, the various method steps described in this application are performed.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

[0153] Furthermore, those skilled in the art will appreciate that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this application and to form distinct embodiments. For example, in the claims above, any of the claimed embodiments may be used in any combination. The information disclosed in this background section is intended solely to enhance understanding of the overall background of this application and should not be construed as an admission or any implication that such information constitutes prior art known to those skilled in the art.

Claims

1. A device monitoring alarm processing method based on zabbix, It is characterized in that The method comprises: Obtain the monitoring data set of the controlled device through the zabbix-agent deployed on the controlled device; Analyzing whether the controlled device has an abnormality based on the monitoring data set, and when an abnormality exists, outputting alarm information about the controlled device; Receiving an analysis instruction for the alarm information; Calling a preset first analysis model to analyze the warning information and output initial analysis information, wherein the initial analysis information reflects the abnormal meaning represented by the warning information; Acquiring background information related to the alarm information, including: parsing the alarm location and the alarm device from the alarm information, calling a camera matching the alarm device to shoot the scene environment information at the alarm location through the RTSP protocol, and using the scene environment information as the background information; Determining at least one second analysis model to be used based on the initial parsing information and the background information; Converting the initial parsed information and the background information into information to be analyzed that can be understood by the second analysis model, comprising: extracting a plurality of sub-data from the initial parsed information and the background information, each sub-data being matched with one or more second analysis models, and converting each sub-data into sub-information to be analyzed that can be understood by the matching second analysis model based on an analysis template of the second analysis model; The second analysis model is called to parse the information to be analyzed, and the analysis result information corresponding to the alarm information is output, wherein the first analysis model is an analysis model that integrates the knowledge base information related to the controlled device and the zabbix framework, and the second analysis model is a large language model, and the second analysis model includes multiple.

2. The method according to claim 1, It is characterized in that The acquiring of background information related to the alarm information includes: Obtaining basic device information of the controlled device; Obtaining an abnormal time period based on the generation time of the alarm information; The monitoring data generated in the abnormal period and related to the initial parsed information is extracted from the monitoring data set, and the basic device information and the extracted monitoring data are used as the background information.

3. The method according to claim 1, It is characterized in that The acquiring of background information related to the alarm information includes: Acquire historical analysis data of historical alarms related to the alarm information, and use the historical analysis data as the background information.

4. The method according to claim 1, It is characterized in that The calling of the second analysis model to parse the information to be analyzed and outputting analysis result information corresponding to the alarm information includes: Determine a parsing order of each sub-data based on the multiple sub-data; According to the parsing order, the second analysis model corresponding to each sub-data is respectively called to process the sub-information to be analyzed; And using the analysis result sub-information of the previous sub-data among the multiple sub-data having the dependency relationship as the input of the second analysis model corresponding to the subsequent sub-data; The analysis result sub-information output by each second analysis model is fused to form the analysis result information.

5. The method according to claim 1, It is characterized in that The controlled device includes a door access device; the method further includes: When it is determined based on the analysis result information that the access control device needs to be remotely controlled, acquiring image data captured by a camera associated with the access control device; The image data is identified, and the opening and release of the access control device is remotely controlled based on the identification result.

6. A device monitoring and alarm processing device based on zabbix, It is characterized in that The device comprises: A monitoring data acquisition module is used to obtain the monitoring data set of the controlled device through the zabbix-agent deployed on the controlled device; An alarm identification module is used to analyze whether the controlled device has an abnormality based on the monitoring data set, and output alarm information about the controlled device when an abnormality exists; An analysis instruction acquisition module, used to receive an analysis instruction for the alarm information; An initial analysis module, used to call a preset first analysis model to analyze the alarm information and output initial analysis information, wherein the initial analysis information reflects the abnormal meaning represented by the alarm information; A background information acquisition module, used to acquire background information related to the alarm information; A secondary analysis module, used to convert the initial analysis information and the background information into information to be analyzed that can be understood by a second analysis model; calling the second analysis model to parse the information to be analyzed, and outputting analysis result information corresponding to the alarm information, wherein the first analysis model is an analysis model that integrates the knowledge base information related to the controlled device and the zabbix framework, and the second analysis model is a large language model; The background information acquisition module is further used to parse the alarm location and the alarm device from the alarm information, and to call a camera matching the alarm device to shoot the scene environment information at the alarm location through the RTSP protocol, and use the scene environment information as the background information; The secondary analysis module is also used to determine at least one second analysis model to be used based on the initial analysis information and the background information, disassemble multiple sub-data from the initial analysis information and the background information, each sub-data matches one or more of the second analysis models, and convert each sub-data into sub-information to be analyzed that can be understood by the matching second analysis model based on the analysis template of the second analysis model, and the second analysis models include multiple.

7. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores executable instructions, and when the executable instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 5.

8. An electronic device, It is characterized in that include: one or more processors; A memory for storing one or more programs, which, when executed by the one or more processors, causes the one or more processors to perform the method according to any one of claims 1 to 5.