WEB application availability alarm linkage verification method, system and equipment

Abstract

The invention discloses a WEB application availability alarm linkage verification method, system and device, and relates to the technical field of WEB application availability alarm verification, and the method comprises the following steps: obtaining account data of a service system, pushing the account data to a performance dial test system through an interface, and obtaining a unique identification code of the service system through Hash processing based on the account data of the service system; when the system availability abnormity is triggered, the availability abnormity prediction model is driven based on the unique identification code; performing judgment, optimally selecting a response optimal path based on an improved Dijkstra algorithm according to a judgment result, and sending an availability detection instruction to the front acquisition equipment; obtaining feedback data after detection is completed, combining the output of the availability exception prediction model, constructing a consistency verification matrix, carrying out fusion judgment, and constructing a dynamic alarm triggering strategy; according to the method, a dynamic alarm triggering strategy is formed through unified identification, multi-source prediction verification and a topology perception path scheduling mechanism, and the accuracy and the real-time performance of Web application availability alarm are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of WEB application availability alarm verification, and more particularly to a WEB application availability alarm linkage verification method, system and device. Background Art

[0002] With the continuous development of internet technology and information system architecture, core business systems in many industries are gradually migrating to web application architectures. System deployment methods are becoming increasingly diverse, encompassing various forms such as on-premises, cloud-based, and hybrid deployments. In this context, ensuring the availability of web applications has become a key issue in enterprise IT operations and maintenance. This is especially true in the context of large-scale, multi-node, and cross-environment deployments. The localizability of faults and the accuracy of alerts are directly related to system response timeliness and operational efficiency.

[0003] Existing performance monitoring and availability alert systems often have the following problems: Traditional systems mostly rely on static threshold trigger mechanisms. When business indicators fluctuate or deployment methods change, problems such as insensitive alarms or frequent false alarms are prone to occur. This is particularly prominent in scenarios with dynamic resource scaling and short-term network jitter.

[0004] In practical applications, system deployment topologies are complex, and most monitoring systems lack topology awareness. Detection paths often rely on fixed configurations or broadcast-based multi-transmission methods, making it difficult to select the optimal detection path for different deployment methods. This increases network load and reduces detection efficiency. This present invention proposes a solution to these problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a WEB application availability alarm linkage verification method, system and device. Through unified identification, multi-source prediction verification and topology-aware path scheduling mechanism, a high-reliability consistency scoring model is constructed to form a dynamic alarm triggering strategy, which significantly improves the accuracy and real-time performance of Web application availability alarms.

[0006] To achieve the above object, the present invention provides the following technical solutions: A web application availability alarm linkage verification method comprises the following steps: obtaining business system ledger data and pushing it to a performance dialing system through an interface, obtaining a unique business system identification code based on the business system ledger data through hash processing; when a system availability anomaly is triggered, driving an availability anomaly prediction model based on the unique identification code; making a judgment based on the output of the model, and according to the judgment result, optimizing and selecting the optimal response path based on an improved Dijkstra algorithm and issuing an availability detection instruction to a front-end collection device; obtaining feedback data after the detection is completed, combining it with the output of the availability anomaly prediction model, constructing a consistency verification matrix, performing a fusion judgment, and constructing a dynamic alarm triggering strategy.

[0007] In a preferred embodiment, a unique identification code of the business system is obtained through hash processing based on the business system ledger data, specifically: the business system ledger data is standardized to obtain a structured feature vector, and the business system ledger data includes the affiliated unit, system standard name, system ID, system type, deployment method and operating environment label; the structured feature vector is input into a preset weight distribution matrix to generate a pre-mapped feature vector; the pre-mapped feature vector is subjected to salted hash processing to generate a unique identification code of the business system; and the unique identification code is used as the key value to store the business system ledger data, structured feature vector and pre-mapped feature vector in the ledger mapping table within the system.

[0008] In a preferred embodiment, the availability anomaly prediction model is driven based on a unique identification code, specifically: the business system ID associated with the abnormal event is passed into the performance dialing system as an input parameter; the performance dialing system retrieves the corresponding unique identification code based on the system ID, and uses the identification code as the key index field, calls the maintained system ledger mapping table, obtains time series data and aligns it in chronological order to form a feature matrix; and calls the anomaly prediction model that matches the target system type and deployment method, and uses the feature matrix as input; obtains the current actual availability value and compares it with the model output, constructs an anomaly deviation scoring function, and outputs an anomaly deviation scoring value.

[0009] In a preferred embodiment, judgment is made through the output of the model, and according to the judgment result, the optimal response path is optimized and selected based on the improved Dijkstra algorithm, and an availability detection instruction is issued to the front-end collection device, specifically: if the abnormal deviation score value is greater than the preset abnormal judgment threshold, the current device topology connection structure is obtained according to the system unique identification code; a graph structure is constructed according to the topology connection structure and the front-end detection device set; based on the graph structure, a sub-graph filtering mechanism is introduced into the deployment method to construct a normalized detection graph model for deployment matching; based on the graph model, with the performance dialing system scheduling center node as the source point, detection is performed through the shortest weighted path search strategy to obtain the shortest detection path; according to the shortest detection path, a detection instruction is initiated to the front-end collection device and detection scheduling is executed.

[0010] In a preferred embodiment, feedback data after detection is completed is obtained, combined with the output of the availability anomaly prediction model, a consistency verification matrix is ​​constructed, and a fusion judgment is performed. Specifically, detection feedback data of multiple dimensions is obtained, and the feedback data is encapsulated into a structured detection vector; the detection vector is fused and compared with reference indicators to generate a consistency verification matrix, and the reference indicators include the actual system availability value, the availability prediction value, and the average availability value of the historical operation stable period; according to the consistency deviation of each indicator between the consistency verification matrices, a weighted consistency scoring function is defined.

[0011] In a preferred embodiment, a fusion judgment is performed to construct a dynamic alarm triggering strategy, specifically: the output of the consistency scoring function is compared and analyzed with the preset dynamic alarm threshold; an alarm confirmation label and a suppression label are generated based on the comparison and analysis results; if it is an alarm confirmation label, a formal alarm trigger instruction is returned, and the alarm context information is written into the log audit module; if it is a suppression label, the current round of alarm process is terminated.

[0012] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By introducing a unique identification code mechanism for business systems and an availability anomaly prediction model based on the Transformer structure, efficient and unique identification of system identities and intelligent prediction of availability status are achieved, with significant advantages in real-time performance, accuracy, and scalability. Among them, the identification code is weighted through structured feature fusion and salted hash processing is introduced to ensure that the ledger mapping is collision-resistant and traceable. The model calling mechanism based on deployment type adaptation makes predictions more targeted and accurate, effectively supporting subsequent anomaly scoring, path scheduling, and dynamic alarm logic. Overall, it significantly improves the prediction accuracy and linkage processing capabilities of the performance dialing system under multi-business and multi-environment conditions.

[0013] 2. By introducing a deployment-aware detection path optimization mechanism based on the improved Dijkstra algorithm and combining multi-dimensional detection feedback with model prediction results to construct a consistency verification matrix, efficient verification of system anomalies and dynamic alarm strategy generation are achieved. It has significant advantages such as detection path adaptation, multi-source fusion of feedback indicators, and accurate and reliable alarms. This mechanism not only improves the real-time and accuracy of anomaly detection, but also has closed-loop learning and model iteration capabilities, significantly enhancing the intelligent and adaptive operation and maintenance level of the performance dialing system in complex deployment environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A flowchart of a WEB application availability alarm linkage verification method provided in an embodiment of the present application.

[0015] Figure 2 A schematic diagram of the structure of a WEB application availability alarm linkage verification system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0017] Example 1, Figure 1 A flowchart of a web application availability alarm linkage verification method provided in an embodiment of the present application includes the following steps: S1, obtain the business system ledger data and push it to the performance dialing system through the interface, and obtain the business system unique identification code through hash processing based on the business system ledger data. The business system ledger data includes the affiliated unit, system standard name, system ID, system type, deployment method and operating environment label.

[0018] In this embodiment, the performance dialing system receives business system ledger data pushed by the SG-I6000 through an interface. The interface refers to the system communication channel or protocol implementation used to push ledger information between the SG-I6000 system and the performance dialing system. Its essence is a technical carrier that supports cross-system data exchange.

[0019] Obtain the business system ledger data and push it to the performance test system through the interface. Based on the business system ledger data, a unique identification code for the business system is obtained through hash processing. Specifically: Standardize the business system ledger data to obtain a structured feature vector; Inputting the structured feature vector into a preset weight distribution matrix to generate a pre-mapped feature vector; Perform salted hashing on the pre-mapped feature vector to generate a unique identification code for the business system; The unique identification code is used as the key value to store the business system ledger data, structured feature vectors and pre-mapped feature vectors into the ledger mapping table within the system.

[0020] It should be noted that the ledger mapping table is used for unified calling during subsequent availability prediction model indexing, detection path scheduling, and alarm link backtracing.

[0021] The specific calculation formula of the weight distribution matrix is ​​as follows:

[0022] Where, is the pre-mapped feature vector, The trained weight matrix is ​​used to improve the recognition of system ID, deployment mode and key performance characteristics in the coding. is the structured feature vector.

[0023] The specific calculation formula of the business system unique identification code is as follows:

[0024] Where, It is the unique identification code of the business system. is a secure hash function, is the salt string, The pre-mapped feature vector is serialized into a byte stream in a specified order and encoded using Base64 to form a standard ASCII string for easy input into the hash function.

[0025] It should be noted that the salting mechanism is introduced to prevent the same business system from generating the same hash result when registered multiple times at different times, thereby improving collision resistance. By encoding business system ledger data with multi-dimensional parameters, a highly unique and distinguishable system identification code is generated. This enables the performance testing system to quickly and accurately complete business system identity recognition and data matching in key steps such as subsequent availability anomaly prediction, detection path scheduling, and model invocation, thereby ensuring the consistency, timeliness, and traceability of scheduling logic.

[0026] S2: When a system availability anomaly is triggered, an availability anomaly prediction model is driven based on the unique identification code. The anomaly prediction model is constructed based on a time series prediction network with a Transformer structure.

[0027] In this embodiment, availability anomaly refers to the phenomenon that the business system becomes unreachable, responds severely timed out, or experiences a sudden increase in interface error rate at a certain moment or time period, resulting in a significant decrease or interruption in the ability to provide external services.

[0028] The availability anomaly prediction model is driven by a unique identification code, specifically: The business system ID associated with the abnormal event is passed into the performance dialing system as an input parameter; The performance dialing system retrieves the corresponding unique identification code based on the system ID and uses the identification code as the key index field. It calls the maintained system ledger mapping table to obtain time series data and aligns them in chronological order to form a feature matrix. The time series data includes response time, reachability, port availability, task success rate, CPU and memory usage. and calling an anomaly prediction model that matches the target system type and deployment mode, taking the feature matrix as input, and outputting an availability prediction result; Obtain the current actual availability value and compare it with the model output, build an abnormal deviation scoring function, and output the abnormal deviation scoring value.

[0029] The specific calculation formula of the characteristic matrix is ​​as follows:

[0030] Where, is the feature matrix, For the Features in time The value of the moment, is the feature dimension, is the sequence length.

[0031] The specific calculation formula of the abnormal prediction model is as follows:

[0032] formula, is the availability prediction value, To correspond to the time series prediction model of a specific business system, the model structure can be Transformer, which is matched and called based on the system's unique identification code and has system-specific learning capabilities and generalization performance.

[0033] The abnormal deviation scoring function is specifically calculated as follows:

[0034] Where, is the abnormal deviation score value, is the actual availability value of the system at the current moment, The next moment system availability prediction value output by the abnormal prediction model, is the historical baseline fluctuation standard deviation, which indicates the natural fluctuation range of the availability value of the business system under normal conditions.

[0035] It should be noted that the actual availability value represents the actual service reachability or responsiveness of the business system at the current point in time. It is real-time observation data collected by the system operation monitoring module (such as the SG-I6000) through detection, logs, and interface call results. The predicted availability value represents the availability level that the model predicts the system should achieve at a certain point in the future based on historical operating conditions. It is the expected health status value of the system automatically generated using a time series prediction model (such as Transformer).

[0036] S3, judges based on the output of the model. Based on the judgment result, the improved Dijkstra algorithm is used to optimize the selection of the optimal response path and issue an availability detection instruction to the front-end collection device.

[0037] In this embodiment, when the system anomaly deviation score is high, the performance testing system must quickly and accurately select a detection path with high reachability and detection efficiency to ensure timely verification of the availability anomaly. Therefore, by constructing a weighted detection graph model and leveraging an improved Dijkstra algorithm, the system automatically selects the path set with the lowest detection cost and optimal link stability from multiple candidate paths. This mechanism significantly improves detection accuracy and timeliness, and is a key execution link in the "prediction-verification-alert" closed-loop mechanism.

[0038] The model output is used to make a judgment. Based on the judgment result, the optimal response path is selected based on the improved Dijkstra algorithm and an availability detection instruction is issued to the front-end collection device. Specifically: If the abnormal deviation score is greater than the preset abnormality judgment threshold, the system extracts the front-end detection device set from the business system ledger database according to the system unique identification code, and obtains the current device topology connection structure. The front-end detection device set includes the deployment mode, the network area to which it belongs, and the access channel identifier. The device topology connection structure includes node connectivity, path hop count, bandwidth and delay; Based on the topological connection structure and the set of front-end detection devices, a graph structure G = (V, E, B) is constructed, where V is a set of nodes, including the performance dialing system master node, front-end detection devices, and business system target nodes. Each node represents a logical or physical device, E is a set of edges, indicating the existence of a reachable network path or logical channel, and B is a set of weights. Based on the graph structure, a subgraph filtering mechanism is introduced to the deployment mode to build a normalized detection graph model GD=(V*,E*,B*) for deployment matching. Based on the graph model, the performance dialing system scheduling center node is used as the source point, and the shortest weighted path search strategy is used to perform detection to obtain the shortest detection path; According to the shortest detection path, a detection instruction is initiated to the front-end collection device and detection scheduling is performed. The detection instruction includes a system identifier, a detection type, a detection time window, and a result return format.

[0039] The specific calculation formula of the shortest detection path is as follows:

[0040] Where, is the shortest detection path, is a graph search algorithm, To detect the graph model, The dispatch source point of the probe instruction.

[0041] S4, obtains feedback data after the detection is completed, combines it with the output of the availability anomaly prediction model, builds a consistency verification matrix, performs fusion judgment, and builds a dynamic alarm triggering strategy.

[0042] In this embodiment, by introducing a multi-source detection data fusion and consistency scoring mechanism, closed-loop verification and dynamic alarm screening of system anomaly triggers are achieved, thereby effectively filtering out false alarms caused by prediction deviations, short-term jitter or non-critical faults, ensuring that SG-I6000 only issues formal alarms when the anomaly is highly reliable, thereby improving the accuracy and stability of system operation and maintenance decisions.

[0043] Obtain feedback data after detection, combine it with the output of the availability anomaly prediction model, build a consistency verification matrix, and perform fusion judgment, specifically: Acquire multiple dimensions of detection feedback data from the front-end detection device in real time, including host reachability status A1, service port connection status A2, interface response time A3, network link delay A4, and device packet loss rate A5; Encapsulate the feedback data into a structured detection vector DJ=[A1, A2, A3, A4, A5]; The detection vector DJ is fused and compared with the reference index to generate a consistency verification matrix , the reference index includes the actual availability value of the system , usability prediction value , the average availability value during the historical stable operation period ; According to the consistency deviation of each indicator between the consistency verification matrices, a weighted consistency scoring function is defined.

[0044] The specific calculation formula of the weighted consistency scoring function is as follows:

[0045] Where, is the consistency score, is the actual availability value, Availability prediction value, is the average availability value during the historical stable operation period, For the The weight factor of the feedback data, For the The normalized abnormality scoring function of the feedback data is 、 、 are weight coefficients respectively.

[0046] It should be noted that the consistency verification matrix is ​​used to comprehensively reflect whether the current system state deviates from the normal trend and to verify the multi-source confirmation of the abnormality.

[0047] The specific calculation formula of the standardized abnormality scoring function is as follows:

[0048] Where, For the The normalized abnormality scoring function of the feedback data is For feedback data, is the historical average, is the standard deviation.

[0049] The above-mentioned information is combined and judged to build a dynamic alarm triggering strategy, specifically: Compare the output of the consistency scoring function with the preset dynamic alarm threshold; If the output of the consistency scoring function is greater than or equal to the preset dynamic alarm threshold, an alarm confirmation label is generated, otherwise a suppression label is generated; If it is an alarm confirmation tag, the formal alarm trigger instruction is sent back to the SG-I6000, and the alarm context information (anomaly score, detection indicator vector, model prediction value, trigger path, etc.) is written to the log audit module; If it is a suppression tag, the current alarm process will be terminated.

[0050] For abnormal events that do not trigger alarms, the performance dialing system conducts error tracing analysis on model input parameters, prediction deviations, detection path information, etc., and adds the current event as a new sample to the model training set, periodically optimizing the discrimination ability of the availability prediction model through the federated learning mechanism.

[0051] Example 2, Figure 2 A schematic diagram of the structure of a web application availability alarm linkage verification system provided in an embodiment of the present application includes an identification code generation module, a model prediction module, an optimal path judgment module, and a consistency verification module. There are connections between the modules: The identification code generation module is used to obtain business system ledger data and push it to the performance dialing system through the interface. Based on the business system ledger data, a unique identification code of the business system is obtained through hash processing; A model prediction module is used to drive the availability anomaly prediction model based on the unique identification code when a system availability anomaly is triggered; The optimal path judgment module is used to make judgments based on the output of the model. Based on the judgment results, it optimizes and selects the optimal response path based on the improved Dijkstra algorithm and issues availability detection instructions to the front-end collection equipment; The consistency verification module is used to obtain feedback data after the detection is completed, combine it with the output of the availability anomaly prediction model, build a consistency verification matrix, perform fusion judgment, and build a dynamic alarm trigger strategy.

[0052] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0053] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0054] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0056] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0057] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for verifying the availability of web applications through alarm linkage, characterized in that: The steps include: Obtain business system ledger data and push it to the performance test system through the interface. Based on the business system ledger data, a unique identification code for the business system is obtained through hash processing; When a system availability anomaly is triggered, the availability anomaly prediction model is driven based on the unique identification code; The model output is used to make a judgment. Based on the judgment result, the optimal response path is selected based on the improved Dijkstra algorithm and an availability detection instruction is issued to the front-end collection equipment. Obtain feedback data after detection is completed, combine it with the output of the availability anomaly prediction model, build a consistency verification matrix, perform fusion judgment, and build a dynamic alarm trigger strategy.

2. The WEB application availability alarm linkage verification method according to claim 1, characterized in that: The business system unique identification code is obtained by hashing the business system ledger data, specifically: Standardize the business system ledger data to obtain a structured feature vector. The business system ledger data includes the affiliated unit, system standard name, system ID, system type, deployment method, and operating environment label; Inputting the structured feature vector into a preset weight distribution matrix to generate a pre-mapped feature vector; Perform salted hashing on the pre-mapped feature vector to generate a unique identification code for the business system; The unique identification code is used as the key value to store the business system ledger data, structured feature vectors and pre-mapped feature vectors into the ledger mapping table within the system.

3. The WEB application availability alarm linkage verification method according to claim 1, characterized in that: The unique identification code-based availability anomaly prediction model is specifically: The business system ID associated with the abnormal event is passed into the performance dialing system as an input parameter; The performance dialing system retrieves the corresponding unique identification code based on the system ID and uses the identification code as the key index field. It calls the maintained system ledger mapping table to obtain time series data and aligns it in chronological order to form a feature matrix. and calling an anomaly prediction model that matches the target system type and deployment mode, taking the feature matrix as input; Obtain the current actual availability value and compare it with the model output, build an abnormal deviation scoring function, and output the abnormal deviation scoring value.

4. The WEB application availability alarm linkage verification method according to claim 1, characterized in that: The output of the model is used to make a judgment. Based on the judgment result, the optimal response path is selected based on the improved Dijkstra algorithm and an availability detection instruction is issued to the front-end collection device. Specifically, If the abnormal deviation score value is greater than the preset abnormal judgment threshold, the current device topology connection structure is obtained according to the system unique identification code; Construct a graph structure based on the topological connection structure and the set of front-end detection devices; Based on the graph structure, a subgraph filtering mechanism is introduced to the deployment mode to build a normalized detection graph model for deployment matching. Based on the graph model, the performance dialing system scheduling center node is used as the source point, and the shortest weighted path search strategy is used to perform detection to obtain the shortest detection path; Initiate detection instructions to the front-end collection device based on the shortest detection path and execute detection scheduling.

5. The WEB application availability alarm linkage verification method according to claim 4, characterized in that: The feedback data obtained after the detection is completed is combined with the output of the availability anomaly prediction model to construct a consistency verification matrix and perform fusion judgment, specifically as follows: Acquire detection feedback data of multiple dimensions, and encapsulate the feedback data into a structured detection vector; The detection vector is integrated and compared with reference indicators to generate a consistency verification matrix. The reference indicators include the actual system availability value, the availability prediction value, and the average availability value during the historical stable operation period; According to the consistency deviation of each indicator between the consistency verification matrices, a weighted consistency scoring function is defined.

6. The WEB application availability alarm linkage verification method according to claim 5, characterized in that: The fusion judgment is performed to construct a dynamic alarm triggering strategy, specifically: Compare and analyze the output of the consistency scoring function with the preset dynamic alarm threshold; Generate alarm confirmation tags and suppression tags based on the comparison and analysis results; If it is an alarm confirmation tag, the formal alarm trigger instruction is returned and the alarm context information is written to the log audit module; If it is a suppression tag, the current alarm process will be terminated.

7. The WEB application availability alarm linkage verification method according to claim 6, characterized in that: The specific calculation formula of the abnormal deviation scoring function is as follows: Where, is the abnormal deviation score value, is the actual availability value of the system at the current moment, The next moment system availability prediction value output by the abnormal prediction model, is the standard deviation of historical baseline fluctuation.

8. The WEB application availability alarm linkage verification method according to claim 7, characterized in that: The specific calculation formula of the weighted consistency scoring function is as follows: Where, is the consistency score, is the actual availability value, Availability prediction value, is the average availability value during the historical stable operation period, For the The weight factor of the feedback data, For the The normalized abnormality scoring function of the feedback data is 、 、 are weight coefficients respectively.

9. A system using the WEB application availability alarm linkage verification method according to any one of claims 1 to 8, characterized in that: It includes the identification code generation module, model prediction module, optimal path judgment module and consistency verification module. There are connections between the modules: The identification code generation module is used to obtain business system ledger data and push it to the performance dialing system through the interface. Based on the business system ledger data, a unique identification code of the business system is obtained through hash processing; A model prediction module is used to drive the availability anomaly prediction model based on the unique identification code when a system availability anomaly is triggered; The optimal path judgment module is used to make judgments based on the output of the model. Based on the judgment results, it optimizes and selects the optimal response path based on the improved Dijkstra algorithm and issues availability detection instructions to the front-end collection equipment; The consistency verification module is used to obtain feedback data after the detection is completed, combine it with the output of the availability anomaly prediction model, build a consistency verification matrix, perform fusion judgment, and build a dynamic alarm trigger strategy.

10. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the WEB application availability alarm linkage verification method according to any one of claims 1 to 8 are implemented.

Images