Cloud platform alarm monitoring analysis method and system based on large model, and storage medium

Abstract

The invention discloses a cloud platform alarm monitoring analysis method and system based on a large model and a storage medium, and the method comprises the steps: collecting historical data, inputting the historical data into the large model, extracting potential fault node information and node state information, and then drawing a node state-fault knowledge spectrogram based on an operation and maintenance knowledge spectrogram; real-time alarm information is obtained, feature information in the alarm information is extracted and compared with existing alarms in an alarm database, if the feature information does not exist in the alarm database, potential fault node information and node state information are obtained based on a large model, fault nodes and fault alarm paths are traced based on a state-fault knowledge spectrogram, and the fault nodes and the fault alarm paths are subjected to fault detection. And generating alarm records, fault characteristics, fault analysis and solution recommendation, and updating an alarm database. According to the method, corresponding fault nodes and node state information are obtained based on alarm information association, and then the node state-fault knowledge spectrogram is drawn, so that potential fault node information and node state information output based on a large model can be conveniently compared and traced to specific fault nodes, accurate positioning is realized, and the method has relatively good practicability.

Description

Technical Field

[0001] The present invention belongs to the technical field of cloud platform alarm monitoring, and specifically relates to a cloud platform alarm monitoring and analysis method, system and storage medium based on a large model. Background Art

[0002] With the deepening of enterprise informatization, cloud platforms have become key infrastructure to support the operation of enterprise business. For companies, the stability and efficiency of cloud platforms are directly related to the company's daily operations and external services. As a plus point in the company's cloud platform operation and maintenance maturity assessment, operation and maintenance automation is not only the company's current need, but also a future development trend. Cloud platforms usually contain a large number of complex systems and services. When problems arise, it is crucial to quickly locate and solve the problems. The automatic analysis system can monitor alarms in real time, respond promptly and provide detailed information about the fault, thereby speeding up the troubleshooting process. Automated operation and maintenance can greatly improve operation and maintenance efficiency, reduce human errors, and ensure the stability and continuity of the system. However, this is also what the company currently lacks. In order to make up for this shortcoming, it is necessary to increase investment in operation and maintenance automation, introduce advanced technologies and solutions, and ensure improvements in a short period of time.

[0003] With the acceleration of digital transformation of enterprises, cloud platforms have higher and higher requirements for automated analysis and handling of alarms. However, with the increasing complexity of the system, traditional alarm handling methods can no longer meet the needs of enterprises. The company's cloud platform is a key component of daily operations, but the average of 500 alarms generated per day has become a challenge for operation and maintenance management. Especially in specific scenarios such as cloud platform upgrades or abnormal operation of cloud service components, the number of alarms may increase rapidly in a short period of time, even reaching tens of thousands per day. Such large-scale alarm events are a huge pressure for the operation and maintenance team. Faced with a large amount of alarm information every day, how to quickly and accurately locate the fault node and trace the alarm path is a huge challenge faced by the operation and maintenance team. In order to improve the efficiency of operation and maintenance and reduce the burden on operation and maintenance personnel, it has become an urgent need to introduce automated alarm handling software. Through the analysis and automatic handling capabilities of cloud platform alarms, the efficiency and accuracy of alarm processing can be greatly improved, ensuring the stable and efficient operation of the cloud platform. Summary of the invention

[0004] The purpose of the present invention is to provide a cloud platform alarm monitoring and analysis method, system and storage medium based on a large model, aiming to solve the above-mentioned problems, quickly and accurately trace the fault nodes and mounting alarm paths, and then quickly obtain relevant solution recommendations.

[0005] The present invention is mainly achieved through the following technical solutions: The cloud platform alarm monitoring and analysis method based on the big model includes the following steps: Step S1: Collect historical data, obtain alarm information, fault node information, node status information, operation and maintenance information, and network topology information; draw an operation and maintenance knowledge spectrum based on the operation and maintenance information and network topology information; Step S2: Input historical data into the big model to extract potential fault node information and node status information, and then draw a node status-fault knowledge spectrum based on the operation and maintenance knowledge spectrum; Step S3: Acquire real-time network operation data to update historical data, extract feature information from the alarm information based on the real-time alarm information in the network operation data, and compare it with the existing alarms in the alarm database to determine whether the alarm exists in the alarm database. If not, proceed to step S4; otherwise, retrieve the fault analysis and solution recommendations for the corresponding alarm. Step S4: Input the alarm information into the big model to obtain potential fault node information and node status information, and trace the fault node and fault alarm path based on the state-fault knowledge spectrum, input the fault alarm path into the big model to generate alarm records, fault characteristics, fault analysis, solution recommendations, and store them in the alarm database.

[0006] In order to better implement the present invention, further, in the step S1, when collecting historical data, preprocessing is performed, and then alarm information, fault node information, node status information, operation and maintenance information and network topology information with time series characteristics are obtained; the preprocessing includes the following steps: Step S11: Data cleaning, removing noise and processing missing values ​​and outliers in turn; Step S12: Data standardization, performing data normalization processing.

[0007] In order to better implement the present invention, further, in step S3, it is determined whether the alarm exists in the alarm database based on the similarity.

[0008] The present invention is mainly achieved through the following technical solutions: The cloud platform alarm monitoring and analysis system based on the big model is based on the above method and includes a data collection module, a knowledge spectrum drawing module, a big model module, a traceability module and a query comparison module; the data collection module is used to collect historical data; the knowledge spectrum drawing module is used to draw the operation and maintenance knowledge spectrum and the node status-fault knowledge spectrum in sequence; the traceability module is used to trace the fault nodes and fault alarm paths based on the status-fault knowledge spectrum, and the big model module is used to generate alarm records, fault characteristics, fault analysis, and solution recommendations based on the fault alarm path.

[0009] In order to better implement the present invention, further, a preprocessing module is included, and the preprocessing module is used to perform data cleaning and data standardization on historical data.

[0010] A computer-readable storage medium stores a computer program, which implements the above method when executed by a processor.

[0011] The beneficial effects of the present invention are as follows: The present invention first obtains the corresponding fault node and node status information based on the alarm information association, and then draws the node status-fault knowledge spectrum, which is convenient for comparing and tracing the potential fault node information and node status information output based on the large model to the specific fault node, and realizes accurate positioning, which has good practicality. The present invention draws the node status-fault knowledge spectrum through the operation and maintenance knowledge spectrum, which is more in line with the fault tracing of the current operation and maintenance; and the present invention is based on historical data and large models, and more accurately extracts the data information for drawing the state-fault knowledge spectrum. Then, based on the feature information in the real-time alarm information, the fault node and the fault alarm path can be traced through the alarm database or the large model and the state-fault knowledge spectrum, and then a new alarm record, fault characteristics, fault analysis, and solution recommendation are generated, and the alarm database is updated, realizing the timely tracing and processing of alarms, as well as updating powerful background processing capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 The present invention is a flowchart of the cloud platform alarm monitoring and analysis method based on a large model. DETAILED DESCRIPTION

[0013] Embodiment 1: Cloud platform alarm monitoring and analysis methods based on large models, such as Figure 1 As shown, the following steps are included: Step S1: Collect historical data, obtain alarm information, fault node information, node status information, operation and maintenance information, and network topology information; draw an operation and maintenance knowledge spectrum based on the operation and maintenance information and network topology information; When collecting historical data, preprocessing is performed to obtain alarm information, fault node information, node status information, operation and maintenance information, and network topology information with time series characteristics; preprocessing includes the following steps: Step S11: Data cleaning, removing noise and processing missing values ​​and outliers in turn; Use filtering methods to remove noise points in the data to make the data cleaner and more reliable. Machine learning algorithms (decision trees, support vector machines) can be used to determine outliers, and then standardize or directly remove them. Fill missing values ​​to complete the missing values ​​in the data. Common methods include mean filling, median filling, interpolation, etc.

[0014] Step S12: Data standardization, performing data normalization processing.

[0015] The purpose of data normalization is to process data of different scales and dimensions. Commonly used methods include Min-Max normalization and Z-Score normalization. Among them, Min-Max normalization is to map the data to a specified range, for example, mapping the data to the range of 0~1 or -1~1 for processing.

[0016] Step S2: Input historical data into the large model AIGC to extract potential fault node information and node status information, and then draw a node status-fault knowledge spectrum based on the operation and maintenance knowledge spectrum; Step S3: Acquire real-time network operation data to update historical data, extract feature information from the alarm information based on the real-time alarm information in the network operation data, and compare it with the existing alarms in the alarm database, and determine whether the alarm exists in the alarm database based on similarity. If not, proceed to step S4, otherwise, retrieve the fault analysis and solution recommendation of the corresponding alarm; Step S4: Input the alarm information into the large model AIGC to obtain potential fault node information and node status information, and trace the fault node and fault alarm path based on the state-fault knowledge spectrum, input the fault alarm path into the large model to generate alarm records, fault characteristics, fault analysis, solution recommendations, and store them in the alarm database.

[0017] The present invention draws a node state-fault knowledge spectrum through an operation and maintenance knowledge spectrum, which is more in line with the fault tracing of current operation and maintenance; and the present invention is based on historical data and a large model to more accurately extract data information for drawing the state-fault knowledge spectrum. Furthermore, based on the characteristic information in the real-time alarm information, the fault node and the fault alarm path can be traced through the alarm database or the large model and the state-fault knowledge spectrum, and then new alarm records, fault characteristics, fault analysis, and solution recommendations are generated, and the alarm database is updated, realizing the timely tracing and processing of alarms, as well as updating powerful background processing capabilities, which has good practicality.

[0018] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A cloud platform alarm monitoring and analysis method based on a large model, characterized in that: The following steps are involved: Step S1: Collect historical data to obtain alarm information, fault node information, node status information, operation and maintenance information, and network topology information; Draw an operation and maintenance knowledge graph based on operation and maintenance information and network topology information; Step S2: Input historical data into the big model to extract potential fault node information and node status information, and then draw a node status-fault knowledge spectrum based on the operation and maintenance knowledge spectrum; Step S3: Acquire real-time network operation data to update historical data, extract feature information from the alarm information based on the real-time alarm information in the network operation data, and compare it with the existing alarms in the alarm database to determine whether the alarm exists in the alarm database. If not, proceed to step S4; otherwise, retrieve the fault analysis and solution recommendations for the corresponding alarm. Step S4: Input the alarm information into the big model to obtain potential fault node information and node status information, and trace the fault node and fault alarm path based on the state-fault knowledge spectrum, input the fault alarm path into the big model to generate alarm records, fault characteristics, fault analysis, solution recommendations, and store them in the alarm database.

2. The cloud platform alarm monitoring and analysis method based on a large model according to claim 1 is characterized in that: In the step S1, when collecting historical data, preprocessing is performed, and then alarm information, fault node information, node status information, operation and maintenance information and network topology information with time series characteristics are obtained; The preprocessing includes the following steps: Step S11: Data cleaning, removing noise and processing missing values ​​and outliers in turn; Step S12: Data standardization, performing data normalization processing.

3. The cloud platform alarm monitoring and analysis method based on a large model according to claim 1 is characterized in that: In step S3, it is determined whether the alarm exists in the alarm database based on the similarity.

4. A cloud platform alarm monitoring and analysis system based on a large model, based on the method described in any one of claims 1 to 3, characterized in that: It includes a data collection module, a knowledge spectrum drawing module, a large model module, a tracing module and a query comparison module; the data collection module is used to collect historical data; the knowledge spectrum drawing module is used to draw the operation and maintenance knowledge spectrum and the node status-fault knowledge spectrum in sequence; the tracing module is used to trace the fault nodes and fault alarm paths based on the status-fault knowledge spectrum, and the large model module is used to generate alarm records, fault characteristics, fault analysis, and solution recommendations based on the fault alarm path.

5. The cloud platform alarm monitoring and analysis system based on a large model according to claim 4 is characterized in that: It also includes a preprocessing module, which is used to clean and standardize the historical data.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method described in any one of claims 1 to 3 is implemented.

Images