Cloud platform fault diagnosis method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a cloud platform fault diagnosis method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining error log data of a cloud platform in real time; dividing the error log data into a plurality of pieces of log data to be diagnosed; inputting the to-be-diagnosed log data into a plurality of error log diagnosis models to obtain an error log diagnosis result; inputting the to-be-diagnosed log data and the error log diagnosis results into a plurality of error log diagnosis models, and obtaining voting results of the plurality of error log diagnosis results output by each error log diagnosis model; summarizing the error log diagnosis results output by the plurality of error log diagnosis models based on the voting results of the plurality of error log diagnosis results to obtain the accuracy rate of each error log diagnosis result; and based on the accuracy of each error log diagnosis result, determining a fault diagnosis result of the cloud platform. According to the invention, the accuracy of the cloud platform fault diagnosis result is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a cloud platform fault diagnosis method, device, computer equipment and storage medium. Background Art

[0002] The vigorous development of cloud computing technology has promoted the scale and complexity of cloud platforms, and fault diagnosis has become a key link in maintaining their stability.

[0003] In the related art, fault diagnosis is performed by analyzing error logs through manual experience. However, in the face of increasingly complex system architectures, the efficiency and accuracy of fault diagnosis in this way cannot meet the requirements.

[0004] The integration of artificial intelligence technology has brought a turnaround to the above problems, especially the intelligent agents built through large models have shown great potential in error log analysis. Such intelligent agents can automatically parse error logs, improving the efficiency and accuracy of fault diagnosis.

[0005] However, when using intelligent agents for fault diagnosis in related technologies, when multiple large models constituting the intelligent agent give inconsistent diagnosis results for the same fault, there is no effective mechanism to resolve this contradiction, resulting in low accuracy of the fault diagnosis results. Summary of the invention

[0006] In view of this, the present invention provides a cloud platform fault diagnosis method, apparatus, computer equipment and storage medium to solve the problem in the related technology that when using an intelligent agent for fault diagnosis, when multiple large models constituting the intelligent agent have inconsistent diagnosis results for the same fault, there is no effective mechanism to resolve this contradiction, resulting in low accuracy of the fault diagnosis results.

[0007] In a first aspect, the present invention provides a cloud platform fault diagnosis method, the method comprising:

[0008] Obtain error log data from the cloud platform in real time;

[0009] Dividing the error log data into a plurality of log data to be diagnosed;

[0010] For any log data to be diagnosed, input the log data to be diagnosed into multiple error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model;

[0011] Inputting the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models, and obtaining voting results of the multiple error log diagnosis results output by each error log diagnosis model;

[0012] Based on the voting results of multiple error log diagnosis results output by each error log diagnosis model, the error log diagnosis results output by multiple error log diagnosis models are aggregated to obtain the accuracy of each error log diagnosis result;

[0013] Based on the accuracy of each error log diagnosis result, the fault diagnosis result of the cloud platform is determined.

[0014] The cloud platform fault diagnosis method provided by the present embodiment inputs any log data to be diagnosed into multiple error log diagnosis models to obtain the error log diagnosis result output by each error log diagnosis model, inputs the log data to be diagnosed and the error log diagnosis result output by each error log diagnosis model into multiple error log diagnosis models to obtain the voting result of the multiple error log diagnosis results output by each error log diagnosis model, summarizes the error log diagnosis results output by the multiple error log diagnosis models based on the voting result of the multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result, determines the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result, and in the case where an intelligent agent constructed by multiple large models performs fault diagnosis on the cloud platform, the fault diagnosis result of the cloud platform is determined through the above-mentioned mechanism, which combines the error log diagnosis results output by multiple large models to improve the accuracy of the fault diagnosis result.

[0015] In an optional implementation, the real-time acquisition of error log data of the cloud platform includes:

[0016] Collect operation logs and fault logs from different server nodes in real time;

[0017] The error log data is filtered out from the operation log and the fault log.

[0018] The cloud platform fault diagnosis method provided in this embodiment improves the efficiency of cloud platform fault diagnosis by automatically filtering error log data from operation logs and fault logs.

[0019] In an optional implementation, dividing the error log data into a plurality of log data to be diagnosed includes:

[0020] Obtaining timestamp information of the error log data;

[0021] Based on the timestamp information, the error log data is divided into a plurality of log data to be diagnosed according to a preset time window;

[0022] The segment log data corresponding to the preset time period before other preset time windows except the first preset time window overlap with the segment log data corresponding to the preset time period after the previous preset time window.

[0023] The cloud platform fault diagnosis method provided in this embodiment divides the error log data into multiple log data to be diagnosed by setting a preset time window, performs cloud platform fault diagnosis based on the multiple log data to be diagnosed, and improves the accuracy of cloud platform fault diagnosis results.

[0024] In some optional implementations, inputting the log data to be diagnosed into multiple error log diagnosis models to obtain the error log diagnosis result output by each error log diagnosis model includes:

[0025] For any error log diagnosis model, obtaining a fault diagnosis template of the error log diagnosis model;

[0026] Inputting the to-be-diagnosed log data into the fault diagnosis template to obtain an instantiated fault diagnosis template;

[0027] The instantiated fault diagnosis template is input into the corresponding error log diagnosis model to obtain the error log diagnosis result output by the error log diagnosis model.

[0028] The cloud platform fault diagnosis method provided in this embodiment instantiates the fault diagnosis template of the error log diagnosis model by using the log data to be diagnosed, so that the error log diagnosis model can perform fault diagnosis on the instantiated fault diagnosis template, thereby ensuring the accuracy of the error log diagnosis result.

[0029] In some optional implementations, inputting the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model includes:

[0030] For any error log diagnosis model, obtaining a diagnosis result voting template of the error log diagnosis model;

[0031] Inputting the to-be-diagnosed log data and the error log diagnosis results output by each error log diagnosis model into the diagnosis result voting template to obtain an instantiated diagnosis result voting template;

[0032] Inputting the instantiated diagnosis result voting template into the corresponding error log diagnosis model to obtain a plurality of error log diagnosis results output by the error log diagnosis model and sorted in descending order according to correctness;

[0033] The voting results of the plurality of error log diagnosis results include a plurality of error log diagnosis results sorted from high to low in terms of correctness.

[0034] The cloud platform fault diagnosis method provided in this embodiment instantiates the diagnosis result voting template of the error log diagnosis model by utilizing the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model, so that the error log diagnosis model can sort the error log diagnosis results output by each error log diagnosis model for correctness, thereby obtaining the accuracy of each error log diagnosis result, determining the fault diagnosis result of the cloud platform, and ensuring the accuracy of the cloud platform fault diagnosis.

[0035] In some optional implementations, the voting results of the multiple error log diagnosis results include multiple error log diagnosis results sorted from high to low in terms of correctness, and the voting results of the multiple error log diagnosis results output by each error log diagnosis model are aggregated to obtain the accuracy of each error log diagnosis result, including:

[0036] Based on the multiple error log diagnosis results output by each error log diagnosis model and sorted from high to low in terms of correctness, determine the correctness scores of the multiple error log diagnosis results output by each error log diagnosis model, wherein an error log diagnosis result with high correctness has a high corresponding correctness score;

[0037] For any error log diagnosis model, based on the historical diagnosis accuracy of the error log diagnosis model, determine the weight of the error log diagnosis result output by the error log diagnosis model, wherein the higher the historical diagnosis accuracy, the greater the weight;

[0038] For any error log diagnosis result, obtain the correctness score and weight corresponding to the error log diagnosis result under each error log diagnosis model;

[0039] For any error log diagnosis model, based on the weight corresponding to the error log diagnosis result under the error log diagnosis model, weighted processing is performed on the correctness score corresponding to the error log diagnosis result under the error log diagnosis model to obtain the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model;

[0040] Summing target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models to obtain the accuracy of the error log diagnosis results;

[0041] Among them, the sum of the weights of the error log diagnosis results output by multiple error log diagnosis models is a preset value.

[0042] The cloud platform fault diagnosis method provided in this embodiment determines the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model by determining the correctness scores of multiple error log diagnosis results output by each error log diagnosis model and determining the weight of the error log diagnosis results output by the error log diagnosis model. The target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models are summed to obtain the accuracy of the error log diagnosis results, and then the fault diagnosis results of the cloud platform are determined. By combining the error log diagnosis results output by multiple large models, the accuracy of the fault diagnosis results of the cloud platform is improved.

[0043] In some optional implementations, determining the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result includes:

[0044] Based on the accuracy of each error log diagnosis result, filter out the error log diagnosis result with the highest accuracy;

[0045] Determine whether the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold;

[0046] When the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold, an alarm is issued, and the error log diagnosis result with the highest accuracy is fed back as a fault diagnosis result of the cloud platform;

[0047] When the accuracy of the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, the error log diagnosis result with the highest accuracy is fed back to the maintenance personnel, so that the maintenance personnel can determine the fault diagnosis result of the cloud platform based on the error log diagnosis result with the highest accuracy.

[0048] The cloud platform fault diagnosis method provided in this embodiment sets a preset accuracy threshold, and compares the accuracy of the error log diagnosis result with the highest accuracy with the preset accuracy threshold. When the accuracy of the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, maintenance personnel access the method to determine the fault diagnosis result of the cloud platform. When the accuracy of the error log diagnosis result with the highest accuracy is greater than the preset accuracy threshold, the error log diagnosis result with the highest accuracy is used as the fault diagnosis result of the cloud platform, thereby ensuring the accuracy of the cloud platform fault diagnosis.

[0049] In a second aspect, the present invention provides a cloud platform fault diagnosis device, the device comprising:

[0050] The first acquisition module is used to obtain the error log data of the cloud platform in real time;

[0051] A division module, used for dividing the error log data into a plurality of log data to be diagnosed;

[0052] A second acquisition module is used for inputting any log data to be diagnosed into multiple error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model;

[0053] A third acquisition module is used to input the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model;

[0054] A fourth acquisition module is used to summarize the error log diagnosis results output by multiple error log diagnosis models based on the voting results of multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result;

[0055] The determination module is used to determine the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result.

[0056] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, computer instructions being stored in the memory, and the processor executing the cloud platform fault diagnosis method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0057] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the cloud platform fault diagnosis method of the first aspect or any corresponding embodiment thereof.

[0058] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions, which are used to enable a computer to execute the cloud platform fault diagnosis method of the above-mentioned first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 is a flow chart of a cloud platform fault diagnosis method according to an embodiment of the present invention;

[0061] Figure 2 is a flowchart of another cloud platform fault diagnosis method according to an embodiment of the present invention;

[0062] Figure 3 is a schematic diagram of a process for obtaining a fault diagnosis result according to an embodiment of the present invention;

[0063] Figure 4 is a schematic diagram of a process for obtaining voting results according to an embodiment of the present invention;

[0064] Figure 5 is a flow chart of another cloud platform fault diagnosis method according to an embodiment of the present invention;

[0065] Figure 6 is a schematic diagram of the structure of a cloud platform fault diagnosis system according to an embodiment of the present invention;

[0066] Figure 7 is a structural block diagram of a cloud platform fault diagnosis device according to an embodiment of the present invention;

[0067] Figure 8 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0069] The continuous development of cloud computing technology has led to the increasing scale and complexity of cloud platforms. Fault diagnosis has become the key to ensuring the stable operation of cloud platforms.

[0070] As the complexity of cloud platforms increases, the stability and reliability of the system face severe challenges. Error logs are an important basis for cloud platform fault diagnosis. In related technologies, the method of analyzing error logs based on manual experience is used for cloud platform fault diagnosis. This method is inefficient and inaccurate. In addition, it is difficult to accurately locate the problem when facing complex system failures.

[0071] The development of artificial intelligence technology has brought a turning point for cloud platform fault diagnosis. The use of large models to build intelligent agents for error log diagnosis shows great potential. The intelligent agents built by a single large model may also obtain inconsistent results for the diagnosis of the same fault, resulting in instability in fault diagnosis. The collaboration of multiple large models can simulate the interaction and evolution process between complex systems and has high adaptability and robustness. Among them, the large model is a large language model, which is a machine learning model with a large parameter scale and complexity. In the field of deep learning, a large model usually refers to a neural network model with millions to billions of parameters. An agent (Artificial Intelligence, referred to as AI) is a machine or software created by humans that can perform various tasks and is designed to imitate human intelligent behavior, such as learning, reasoning, perception, understanding language, and recognizing patterns. AI systems are designed to act independently or in collaboration with humans to solve problems and perform tasks.

[0072] However, when using intelligent agents for fault diagnosis in related technologies, when multiple large models constituting the intelligent agent have inconsistent diagnosis results for the same fault, there is no effective mechanism to resolve this contradiction, resulting in low accuracy of fault diagnosis.

[0073] An embodiment of the present invention provides a cloud platform fault diagnosis method, which includes inputting any log data to be diagnosed into multiple error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model, inputting the log data to be diagnosed and the error log diagnosis result output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model, and summarizing the error log diagnosis results output by the multiple error log diagnosis models based on the voting results of the multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result, and determining the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result to achieve the effect of improving the accuracy of the fault diagnosis result and improving the efficiency of the fault diagnosis.

[0074] According to an embodiment of the present invention, an embodiment of a cloud platform fault diagnosis method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0075] In this embodiment, a cloud platform fault diagnosis method is provided, which can be used for a cloud platform. Figure 1 is a flow chart of a cloud platform fault diagnosis method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0076] Step S101, obtaining error log data of the cloud platform in real time.

[0077] The cloud platform constructs a cloud platform fault diagnosis agent through multiple large models. The log receiving module of the cloud platform fault diagnosis agent collects all error log data in real time, including system error log data, cloud platform error log data, application error log data, etc.

[0078] Step S102: divide the error log data into a plurality of log data to be diagnosed.

[0079] After the error log data of the cloud platform is obtained, it is parsed and structured according to the format and content of the error log data for subsequent analysis and query, and the error log data is divided into a plurality of log data to be diagnosed.

[0080] That is to say, the error log data is sliced ​​to obtain a plurality of log data to be diagnosed.

[0081] Step S103: for any log data to be diagnosed, input the log data to be diagnosed into a plurality of error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model.

[0082] Among them, for any log data to be diagnosed, the log data to be diagnosed is input into multiple error log diagnosis models in the cloud platform fault diagnosis agent to obtain the error log diagnosis result output by each error log diagnosis model.

[0083] It should be noted that the error log diagnosis model is a large model, and the multiple error log diagnosis models in the cloud platform fault diagnosis agent are multiple large models that constitute the diagnosis agent.

[0084] The embodiment of the present invention selects a suitable big model for the error log diagnosis scenario to build a cloud platform fault diagnosis agent. Three types of big models are usually used: open source general big models, such as the llama series, open source coding big models, such as codellama, wizardcoder, etc., and self-training big models. These three big models are downloaded to the cloud environment to build a cloud platform fault diagnosis agent. For cloud environments that do not require much Internet security, the big model can also be called by directly calling the third-party big model API service to build a cloud platform fault diagnosis agent. Among them, the self-training big model is a big model obtained by fine-tuning instructions based on the open source big model and its own logs and knowledge.

[0085] Step S104: input the log data to be diagnosed and the error log diagnosis result output by each error log diagnosis model into multiple error log diagnosis models to obtain a voting result of the multiple error log diagnosis results output by each error log diagnosis model.

[0086] After obtaining the error log diagnosis results output by each error log diagnosis model for the log data to be diagnosed, the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model for the log data to be diagnosed are input into multiple error log diagnosis models of the cloud platform fault diagnosis agent to obtain the voting results of the multiple error log diagnosis results output by each error log diagnosis model.

[0087] Exemplarily, the cloud platform fault diagnosis agent includes three error log diagnosis models. After obtaining the error log diagnosis results output by each error log diagnosis model for the log data to be diagnosed, the log data to be diagnosed and the three error log diagnosis results are respectively input into the three error log diagnosis models to obtain the voting results of the three error log diagnosis results output by each error log diagnosis model.

[0088] Step S105 , based on the voting results of the multiple error log diagnosis results output by each error log diagnosis model, the error log diagnosis results output by the multiple error log diagnosis models are aggregated to obtain the accuracy of each error log diagnosis result.

[0089] Among them, after obtaining the voting results of multiple error log diagnosis results output by each error log diagnosis model, the error log diagnosis results output by multiple error log diagnosis models are summarized according to the voting results of the multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result.

[0090] Step S106, determining the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result.

[0091] After obtaining the accuracy of each error log diagnosis result, the fault diagnosis result of the cloud platform is determined according to the accuracy of each error log diagnosis result.

[0092] The cloud platform fault diagnosis method provided by the present embodiment inputs any log data to be diagnosed into multiple error log diagnosis models to obtain the error log diagnosis result output by each error log diagnosis model, inputs the log data to be diagnosed and the error log diagnosis result output by each error log diagnosis model into multiple error log diagnosis models to obtain the voting result of the multiple error log diagnosis results output by each error log diagnosis model, summarizes the error log diagnosis results output by the multiple error log diagnosis models based on the voting result of the multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result, determines the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result, and in the case where an intelligent agent constructed by multiple large models performs fault diagnosis on the cloud platform, the fault diagnosis result of the cloud platform is determined through the above-mentioned mechanism, which combines the error log diagnosis results output by multiple large models to improve the accuracy of the fault diagnosis result.

[0093] The cloud platform fault diagnosis method provided in this embodiment constructs a cloud platform fault diagnosis intelligent entity through multiple large models. These large models share error log data and diagnose each section of log data to be diagnosed. Then, each large model votes on the diagnosis results of the same section of log data to be diagnosed, and finally forms a unified diagnosis result, which ensures the uniqueness of the fault diagnosis result and improves the accuracy of the fault diagnosis result.

[0094] In this embodiment, a cloud platform fault diagnosis method is provided, which can be used for a cloud platform. Figure 2 is a flow chart of a cloud platform fault diagnosis method according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0095] Step S201, obtaining error log data of the cloud platform in real time.

[0096] Specifically, the above step S201 includes:

[0097] Step S2011, collecting operation logs and fault logs from different server nodes in real time.

[0098] The log receiving module is used to receive the operation logs and fault logs collected in real time from different server nodes by collection tools (such as Fluentd and Logstash).

[0099] Step S2012, filter out error log data from the operation log and the fault log.

[0100] Among them, error log data extraction is performed for different types of logs, namely, operation logs and fault logs, that is, error log data is screened out from the operation logs and fault logs.

[0101] It should be noted that after filtering out the error log data, data cleaning and formatting are performed, and noise data is removed and the log format is standardized.

[0102] At the same time, the log receiving module temporarily stores the received error log data through the cache to ensure the reliable collection and processing of error log data.

[0103] Step S202: divide the error log data into a plurality of log data to be diagnosed.

[0104] Specifically, the above step S202 includes:

[0105] Step S2021, obtaining timestamp information of error log data.

[0106] Among them, after collecting operation logs and fault logs from different server nodes, for different types of logs, key information such as timestamps are extracted according to the log type, and the timestamps of all logs are converted into a unified format to ensure the consistency of subsequent log processing.

[0107] After filtering out the error log data, obtain timestamp information of the error log data.

[0108] Step S2022: based on the timestamp information, divide the error log data into a plurality of log data to be diagnosed according to a preset time window.

[0109] The size of the preset time window can be set by a technician. Exemplarily, the size of the preset time window is 5 minutes.

[0110] Formulate a suitable sharding strategy, that is, divide the error log data into multiple log data to be diagnosed according to the preset time window according to the time sequence and the timestamp information of the error log data. It should be noted that in order to ensure the integrity of the shards, each log data to be diagnosed can overlap slightly.

[0111] That is, the segment log data corresponding to the preset time period before other preset time windows except the first preset time window overlaps with the segment log data corresponding to the preset time period after the previous preset time window. It should be noted that the preset time period is set by the technician, and generally, the preset time period can be set to 1 minute.

[0112] It is understandable that the overlapping time length of each preset time window is 1 minute.

[0113] This embodiment adopts a sliding window mechanism, and each time window contains part of the data of the previous window at the beginning to ensure that the data between time windows overlap. The step size of the time window is equal to the time window size minus the overlapping time length.

[0114] After generating a plurality of log data to be diagnosed, a label is generated for each log data to be diagnosed, marking its time window and overlapping part.

[0115] It should be noted that each log data to be diagnosed is assigned one or more fault diagnosis tasks, and the goal of the fault diagnosis task is to analyze the log data to be diagnosed and identify any possible faults or problems.

[0116] Step S203: for any log data to be diagnosed, input the log data to be diagnosed into multiple error log diagnosis models to obtain the error log diagnosis result output by each error log diagnosis model.

[0117] Specifically, the above step S203 includes:

[0118] Step S2031: for any error log diagnosis model, obtain a fault diagnosis template of the error log diagnosis model.

[0119] Each error log diagnosis model is provided with a different fault diagnosis template. For any error log diagnosis model, a fault diagnosis template corresponding to the error log diagnosis model is obtained. The fault diagnosis template is a fault diagnosis prompt template.

[0120] Step S2032: input the log data to be diagnosed into the fault diagnosis template to obtain an instantiated fault diagnosis template.

[0121] Figure 3 FIG. 1 is a flowchart of obtaining fault diagnosis results according to an embodiment of the present invention. Figure 3 As shown, the shard log is the log data to be diagnosed. This embodiment is described by taking three error log diagnosis models to form a cloud platform fault diagnosis agent as an example, and the three error log diagnosis models are model 1, model 2, and model 3.

[0122] After obtaining the fault diagnosis template corresponding to the error log diagnosis model, the log data to be diagnosed is brought into the fault diagnosis template corresponding to the error log diagnosis model to implement instantiation of the fault diagnosis template.

[0123] Step S2033: input the instantiated fault diagnosis template into the corresponding error log diagnosis model to obtain the error log diagnosis result output by the error log diagnosis model.

[0124] Among them, after obtaining the instantiated fault diagnosis template, the instantiated fault diagnosis template is input into the corresponding error log diagnosis model to obtain the error log diagnosis result output by the error log diagnosis model, and then the error log diagnosis result output by each error log diagnosis model for the log data to be diagnosed can be obtained.

[0125] Each error log diagnosis model, i.e., each large model, will analyze the log data to be diagnosed according to its own algorithm and model parameters, and diagnose the fault diagnosis result of the cloud platform based on its own analysis result.

[0126] Exemplarily, the fault diagnosis template can be as follows:

[0127] """

[0128] <Instruction>You are a senior software engineer and you need to help me locate various platform faults and analyze the root causes. Answer the questions concisely and professionally. If you are not sure, please say "Unable to answer this question"< / Instruction>

[0129] <Question>The log information is as follows {{log}}, and there may be multiple faults in the log. Please give the first-occurring fault information, analyze the cause of the fault, and give a repair suggestion< / Question>

[0130] """

[0131] log is the log in the current time window, and the prompt template needs to be instantiated through the log.

[0132] In step S204, the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model are input into multiple error log diagnosis models to obtain the voting result of the multiple error log diagnosis results output by each error log diagnosis model. For details, please refer to Figure 1 Step S104 of the illustrated embodiment, which will not be elaborated here.

[0133] In step S205, based on the voting result of the multiple error log diagnosis results output by each error log diagnosis model, the error log diagnosis results output by the multiple error log diagnosis models are summarized to obtain the accuracy rate of each error log diagnosis result. For details, please refer to Figure 1 Step S105 of the illustrated embodiment, which will not be elaborated here.

[0134] In step S206, based on the accuracy rate of each error log diagnosis result, the fault diagnosis result of the cloud platform is determined. For details, please refer to Figure 1 Step S106 of the illustrated embodiment, which will not be elaborated here.

[0135] The cloud platform fault diagnosis method provided in this embodiment improves the efficiency of cloud platform fault diagnosis by automatically screening error log data from operation logs and fault logs.

[0136] By setting a preset time window, the error log data is divided into multiple log data to be diagnosed, and the fault diagnosis of the cloud platform is performed according to the multiple log data to be diagnosed, improving the accuracy of the cloud platform fault diagnosis result.

[0137] By using the log data to be diagnosed to instantiate the fault diagnosis template of the error log diagnosis model, the error log diagnosis model can perform fault diagnosis on the instantiated fault diagnosis template, thereby ensuring the accuracy of the error log diagnosis result.

[0138] In some optional implementations, the above step S204 includes:

[0139] Step a1: for any error log diagnosis model, obtain a diagnosis result voting template of the error log diagnosis model.

[0140] Different error log diagnosis models are provided with different diagnosis result voting templates. For any error log diagnosis model, the diagnosis result voting template of the error log diagnosis model is obtained, that is, the diagnosis result prompt voting template.

[0141] Step a2: input the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into the diagnosis result voting template to obtain an instantiated diagnosis result voting template.

[0142] The log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model for the log data to be diagnosed are input into the diagnosis result voting template to obtain an instantiated diagnosis result voting template.

[0143] Step a3: input the instantiated diagnosis result voting template into the corresponding error log diagnosis model to obtain multiple error log diagnosis results output by the error log diagnosis model and sorted in descending order according to correctness.

[0144] The voting result of the multiple error log diagnosis results includes the multiple error log diagnosis results sorted from high to low in terms of correctness.

[0145] Figure 4 FIG. 1 is a flow chart of obtaining voting results according to an embodiment of the present invention. Figure 4 As shown, after obtaining the instantiated diagnosis result voting template, the instantiated diagnosis result voting template is input into the corresponding error log diagnosis model to obtain multiple error log diagnosis results output by the error log diagnosis model and sorted in descending order according to correctness, and then multiple error log diagnosis results output by each error log diagnosis model and sorted in descending order according to correctness can be obtained.

[0146] Exemplarily, the diagnosis result voting template may be as follows:

[0147] """

[0148] <Instruction>You are a senior software engineer. Analyze the following diagnostic results based on the logs and answer the questions concisely and professionally. The format is as follows: "Model 1; Model 2", and no other output is allowed.< / Instruction>

[0149] <Question>The log information is as follows {{log}}, Model 1 {{Diagnosis 1}}, Model 2 {{Diagnosis 2}}, Model 3 {{Diagnosis 3}}. Please determine whether the diagnostic results of the three models are correct, output the correct diagnostic results and their sorting, and do not output the models with inaccurate results.< / Question>

[0150] """

[0151] It should be noted that the diagnostic summary includes summarizing the log data to be diagnosed and the error log diagnostic results output by each error log diagnostic model.

[0152] For the error log diagnostic results output by each error log diagnostic model for the log data to be diagnosed, a consensus mechanism is used for large model voting to obtain the voting results of the multiple error log diagnostic results output by each error log diagnostic model.

[0153] The cloud platform fault diagnosis method provided in this embodiment instantiates the diagnostic result voting template of the error log diagnostic model by using the log data to be diagnosed and the error log diagnostic results output by each error log diagnostic model, enabling the error log diagnostic model to sort the error log diagnostic results output by each error log diagnostic model according to correctness, and then obtaining the accuracy rate of each error log diagnostic result, determining the fault diagnosis result of the cloud platform, and ensuring the accuracy of the cloud platform fault diagnosis.

[0154] In some alternative embodiments, the above step S205 includes:

[0155] Step b1, based on the multiple error log diagnostic results output by each error log diagnostic model and sorted from high to low according to correctness, determine the correctness scores of the multiple error log diagnostic results output by each error log diagnostic model, where the error log diagnostic results with higher correctness correspond to higher correctness scores.

[0156] Suppose there are n error log diagnostic models. For any error log diagnostic model, obtain the n error log diagnostic results output by it and sorted from high to low according to correctness, and assign corresponding correctness scores to the n error log diagnostic results according to the correctness level of the n error log diagnostic results.

[0157] Exemplarily, the error log diagnostic result with the first correctness sorting has a correctness score of 1 point, the error log diagnostic result with the second correctness sorting has a correctness score of 0.8 points, and the error log diagnostic results not included have a correctness score of 0 points.

[0158] Step b2, for any error log diagnosis model, based on the historical diagnosis accuracy of the error log diagnosis model, determine the weight of the error log diagnosis result output by the error log diagnosis model, wherein the higher the historical diagnosis accuracy, the greater the weight.

[0159] The weight of each error log diagnosis model is dynamically adjusted according to its historical diagnosis accuracy, which means that error log diagnosis models that performed better in the past will have greater weights.

[0160] Assuming there are n error log diagnosis models, a weight Wi is assigned to each error log diagnosis model Mi. The weight assignment can be based on the historical diagnosis accuracy of the error log diagnosis model, etc., and can be manually modified. Among them, the sum of the weights should be 1.

[0161] Step b3: for any error log diagnosis result, obtain the correctness score and weight corresponding to the error log diagnosis result under each error log diagnosis model.

[0162] For any error log diagnosis result among multiple error log diagnosis results output by multiple error log diagnosis models, obtain the correctness score and weight corresponding to the error log diagnosis result under each error log diagnosis model.

[0163] Step b4, for any error log diagnosis model, based on the weight corresponding to the error log diagnosis result under the error log diagnosis model, weighted processing is performed on the correctness score corresponding to the error log diagnosis result under the error log diagnosis model to obtain the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model.

[0164] Among them, for any error log diagnosis model, based on the weight corresponding to the error log diagnosis result under the error log diagnosis model, the correctness score corresponding to the error log diagnosis result under the error log diagnosis model is weighted, including: multiplying the weight corresponding to the error log diagnosis result under the error log diagnosis model and the correctness score corresponding to the error log diagnosis result under the error log diagnosis model to obtain the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model, and then obtaining the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models.

[0165] Step b5, summing up the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models to obtain the accuracy of the error log diagnosis results.

[0166] Among them, through the above method, the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models are obtained, and the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models are summed to obtain the accuracy of the error log diagnosis results.

[0167] Specifically, summing up the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models includes: Among them, Ri is the target correctness score corresponding to the error log diagnosis result under the i-th error log diagnosis model.

[0168] The sum of the weights of the error log diagnosis results output by the multiple error log diagnosis models is a preset value, which is 1.

[0169] like Figure 4 As shown, the voting aggregation is to aggregate the error log diagnosis results output by multiple error log diagnosis models based on the voting results of the multiple error log diagnosis results output by each error log diagnosis model, including the above steps b1 to b5.

[0170] The cloud platform fault diagnosis method provided in this embodiment determines the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model by determining the correctness scores of multiple error log diagnosis results output by each error log diagnosis model and determining the weight of the error log diagnosis results output by the error log diagnosis model. The target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models are summed to obtain the accuracy of the error log diagnosis results, and then the fault diagnosis results of the cloud platform are determined. By combining the error log diagnosis results output by multiple large models, the accuracy of the fault diagnosis results of the cloud platform is improved.

[0171] In some optional implementations, the above step S206 includes:

[0172] Step c1, based on the accuracy of each error log diagnosis result, filter out the error log diagnosis result with the highest accuracy.

[0173] After the accuracy of each error log diagnosis result is obtained, the error log diagnosis result with the highest accuracy is screened out from the multiple error log diagnosis results.

[0174] The code for obtaining the accuracy of each error log diagnosis result and filtering the error log diagnosis result with the highest accuracy can be as follows:

[0175]

[0176] Step c2, determining whether the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold.

[0177] The preset accuracy threshold is set by the technician based on experience, and generally, the preset accuracy threshold is 0.8. It should be noted that the preset accuracy threshold needs to be adjusted according to the actual situation of the cloud platform to avoid misdiagnosis and potential risks.

[0178] Step c3: when the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold, an alarm is issued, and the error log diagnosis result with the highest accuracy is fed back as the fault diagnosis result of the cloud platform.

[0179] Among them, the accuracy of the error log diagnosis result with the highest accuracy is greater than the preset accuracy threshold, indicating that the error log diagnosis result is accurate enough, a cloud platform failure occurs, and an alarm needs to be issued. The error log diagnosis result with the highest accuracy is the fault diagnosis result of the cloud platform, and the error log diagnosis result with the highest accuracy is fed back to the maintenance personnel of the cloud platform.

[0180] Step c4, when the accuracy of the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, the error log diagnosis result with the highest accuracy is fed back to the maintenance personnel, so that the maintenance personnel can determine the fault diagnosis result of the cloud platform based on the error log diagnosis result with the highest accuracy.

[0181] If the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, it means that the error log diagnosis result is not accurate enough. Then the error log diagnosis result with the highest accuracy is fed back to the maintenance personnel so that the maintenance personnel can determine the fault diagnosis result of the cloud platform based on the error log diagnosis result with the highest accuracy.

[0182] It should be noted that the final result is obtained based on the voting summary, that is, based on the accuracy of each error log diagnosis result, to determine the fault diagnosis result of the cloud platform.

[0183] The cloud platform fault diagnosis method provided in this embodiment sets a preset accuracy threshold, and compares the accuracy of the error log diagnosis result with the highest accuracy with the preset accuracy threshold. When the accuracy of the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, maintenance personnel access the method to determine the fault diagnosis result of the cloud platform. When the accuracy of the error log diagnosis result with the highest accuracy is greater than the preset accuracy threshold, the error log diagnosis result with the highest accuracy is used as the fault diagnosis result of the cloud platform, thereby ensuring the accuracy of the cloud platform fault diagnosis.

[0184] In some optional implementations, the cloud platform fault diagnosis method further includes:

[0185] When the cloud platform stability requirement is the first requirement, the preset accuracy threshold is set as the first threshold.

[0186] When the cloud platform stability requirement is the second requirement, the preset accuracy threshold is set as the second threshold.

[0187] The first required stability level is greater than the second required stability level, and the first threshold is greater than the second threshold.

[0188] That is, during the period when the stability of the cloud platform is required to be high, the preset accuracy threshold is set to a higher threshold to reduce false alarms of faults. During the regular maintenance period, that is, the period when the stability of the cloud platform is not required to be high, more potential problems and faults are captured by setting the preset accuracy threshold to a lower threshold.

[0189] The cloud platform fault diagnosis method provided in this embodiment can better adapt to different operating conditions and maintenance needs by dynamically adjusting the preset accuracy threshold according to the different stability requirements of the cloud platform. It ensures accurate monitoring during high-stability periods and improves the sensitivity of problem discovery in maintenance devices, effectively improving the overall operation and maintenance efficiency and user experience of the cloud platform.

[0190] In this embodiment, a cloud platform fault diagnosis method is also provided. Figure 5 is a flow chart of a cloud platform fault diagnosis method according to an embodiment of the present invention. Figure 5 As shown, the cloud platform fault diagnosis method includes log collection, sharding process, fault diagnosis process, voting process and feedback process.

[0191] Among them, the log collection and sharding process corresponds to the aforementioned step S201 and step S202. Please refer to the aforementioned step S201 and step S202 for detailed description, which will not be repeated here.

[0192] The fault diagnosis process corresponds to the aforementioned step S203 . For detailed description, please refer to the aforementioned step S203 , which will not be repeated here.

[0193] The voting process corresponds to the aforementioned step S204 and step S205. For detailed description, please refer to the aforementioned step S204 and step S205, which will not be repeated here.

[0194] The feedback process corresponds to the aforementioned step S206. For detailed description, please refer to the aforementioned step S206, which will not be repeated here.

[0195] The cloud platform fault diagnosis method provided in this embodiment constructs a cloud platform fault diagnosis agent through multiple large models. These large models share error log data and diagnose each segment of error log data. Then, each large model votes on the diagnosis results for the same segment of error log data, and finally forms a consistent diagnosis result, so that cloud platform faults can be quickly located and maintenance costs can be reduced.

[0196] In this embodiment, a cloud platform fault diagnosis system is also provided. Figure 6 Schematic diagram of the structure of the cloud platform fault diagnosis system according to an embodiment of the present invention. Figure 6 As shown, the cloud platform fault diagnosis system includes a log receiving module, a log sharding distribution module, a large model diagnosis module, a large model voting module and a processing module.

[0197] The log receiving module is used to execute the above step S201, the log shard distribution module is used to execute the above step S202, the large model diagnosis module is used to execute the above step S203, the large model voting module is used to execute the above steps S204 and S205, and the processing module is used to execute the above step S206. For detailed description, please refer to the corresponding part of the above embodiment, which will not be repeated here.

[0198] The cloud platform fault diagnosis system provided in this embodiment integrates multiple large models into one system, and each large model may focus on different types of faults or have different diagnostic algorithms. This collaborative mechanism can improve the accuracy and coverage of cloud platform fault diagnosis. By voting for the multiple diagnostic results obtained by each large model and assigning different weights to each large model, its influence in the final decision can be adjusted according to the reliability of the large model.

[0199] It should be noted that the cloud platform fault diagnosis system provided in this embodiment allows users to customize and integrate a new large model, namely, an error log diagnosis model, according to their own needs, thereby increasing the flexibility and scalability of the system.

[0200] Moreover, the cloud platform fault diagnosis agent in this embodiment is deeply integrated with the cloud platform's log system, and can seamlessly access cloud services and infrastructure, simplifying the deployment and maintenance process.

[0201] In this embodiment, a cloud platform fault diagnosis device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0202] This embodiment provides a cloud platform fault diagnosis device, such as Figure 7 As shown, including:

[0203] The first acquisition module 701 is used to acquire error log data of the cloud platform in real time.

[0204] The division module 702 is used to divide the error log data into a plurality of log data to be diagnosed.

[0205] The second acquisition module 703 is used to input any log data to be diagnosed into multiple error log diagnosis models to obtain the error log diagnosis result output by each error log diagnosis model.

[0206] The third acquisition module 704 is used to input the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model.

[0207] The fourth acquisition module 705 is used to aggregate the error log diagnosis results output by multiple error log diagnosis models based on the voting results of the multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result.

[0208] The determination module 706 is used to determine the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result.

[0209] In some optional implementations, the first acquisition module 701 includes:

[0210] The collection unit is used to collect operation logs and fault logs from different server nodes in real time.

[0211] The first screening unit is used to screen out error log data from the operation log and the fault log.

[0212] In some optional implementations, the partitioning module 702 includes:

[0213] The first acquisition unit is used to acquire timestamp information of the error log data.

[0214] The first division unit is used to divide the error log data into a plurality of log data to be diagnosed according to a preset time window based on the timestamp information.

[0215] The segment log data corresponding to the preset time period before other preset time windows except the first preset time window overlap with the segment log data corresponding to the preset time period after the previous preset time window.

[0216] In some optional implementations, the second acquisition module 703 includes:

[0217] The second acquisition unit is used to acquire the fault diagnosis template of the error log diagnosis model for any error log diagnosis model.

[0218] The first instantiation unit is used to input the log data to be diagnosed into the fault diagnosis template to obtain an instantiated fault diagnosis template.

[0219] The third acquisition unit is used to input the instantiated fault diagnosis template into the corresponding error log diagnosis model to obtain the error log diagnosis result output by the error log diagnosis model.

[0220] In some optional implementations, the third acquisition module 704 includes:

[0221] The fourth acquisition unit is used to acquire the diagnosis result voting template of the error log diagnosis model for any error log diagnosis model.

[0222] The second instantiation unit is used to input the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into the diagnosis result voting template to obtain an instantiated diagnosis result voting template.

[0223] The fifth acquisition unit is used to input the instantiated diagnosis result voting template into the corresponding error log diagnosis model to obtain multiple error log diagnosis results output by the error log diagnosis model and sorted from high to low in terms of correctness.

[0224] The voting result of the multiple error log diagnosis results includes the multiple error log diagnosis results sorted from high to low in terms of correctness.

[0225] In some optional implementations, the fourth acquisition module 705 includes:

[0226] The first determination unit is used to determine the correctness scores of multiple error log diagnosis results output by each error log diagnosis model based on multiple error log diagnosis results output by each error log diagnosis model and sorted from high to low in terms of correctness, wherein an error log diagnosis result with high correctness has a corresponding high correctness score.

[0227] The second determination unit is used to determine the weight of the error log diagnosis result output by the error log diagnosis model for any error log diagnosis model based on the historical diagnosis accuracy of the error log diagnosis model, wherein the higher the historical diagnosis accuracy, the greater the weight.

[0228] The sixth acquisition unit is used to acquire, for any error log diagnosis result, a correctness score and a weight corresponding to the error log diagnosis result under each error log diagnosis model.

[0229] The seventh acquisition unit is used to perform weighted processing on the correctness score of the error log diagnosis result under any error log diagnosis model based on the weight corresponding to the error log diagnosis result under the error log diagnosis model, so as to obtain the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model.

[0230] The eighth acquisition unit is used to sum the target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models to obtain the accuracy of the error log diagnosis results.

[0231] Among them, the sum of the weights of the error log diagnosis results output by multiple error log diagnosis models is a preset value.

[0232] In some optional implementations, the determination module 706 includes:

[0233] The second screening unit is used to screen out the error log diagnosis result with the highest accuracy based on the accuracy of each error log diagnosis result.

[0234] The judgment unit is used to judge whether the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold.

[0235] The alarm unit is used to issue an alarm when the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold, and to feed back the error log diagnosis result with the highest accuracy as the fault diagnosis result of the cloud platform.

[0236] The feedback unit is used to feed back the error log diagnosis result with the highest accuracy to the maintenance personnel when the accuracy of the error log diagnosis result with the highest accuracy is not greater than a preset accuracy threshold, so that the maintenance personnel can determine the fault diagnosis result of the cloud platform based on the error log diagnosis result with the highest accuracy.

[0237] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0238] The cloud platform fault diagnosis device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0239] The embodiment of the present invention also provides a computer device having the above Figure 7 The cloud platform fault diagnosis device shown.

[0240] See also Figure 8 , Figure 8 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 8 As shown, the computer device includes: one or more processors 801, memory 802, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Each component utilizes different buses to communicate with each other, and can be installed on a common mainboard or installed in other ways as required. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 8 A processor 801 is taken as an example.

[0241] The processor 801 may be a central processing unit, a network processor or a combination thereof. The processor 801 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable logic gate array, a general purpose array logic or any combination thereof.

[0242] The memory 802 stores instructions executable by at least one processor 801, so that the at least one processor 801 executes the method shown in the above embodiment.

[0243] The memory 802 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 802 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 802 may optionally include a memory remotely arranged relative to the processor 801, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0244] The memory 802 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 802 may also include a combination of the above types of memory.

[0245] The computer device also includes a communication interface 803, which is used for the computer device to communicate with other devices or a communication network.

[0246] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0247] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0248] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A cloud platform fault diagnosis method, characterized in that: The method comprises: Obtain error log data from the cloud platform in real time; Dividing the error log data into a plurality of log data to be diagnosed; For any log data to be diagnosed, input the log data to be diagnosed into multiple error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model; Inputting the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models, and obtaining voting results of the multiple error log diagnosis results output by each error log diagnosis model; Based on the voting results of multiple error log diagnosis results output by each error log diagnosis model, the error log diagnosis results output by multiple error log diagnosis models are aggregated to obtain the accuracy of each error log diagnosis result; Based on the accuracy of each error log diagnosis result, the fault diagnosis result of the cloud platform is determined.

2. The method according to claim 1, characterized in that The real-time acquisition of error log data of the cloud platform includes: Collect operation logs and fault logs from different server nodes in real time; The error log data is filtered out from the operation log and the fault log.

3. The method according to claim 1, characterized in that The step of dividing the error log data into a plurality of log data to be diagnosed includes: Obtaining timestamp information of the error log data; Based on the timestamp information, the error log data is divided into a plurality of log data to be diagnosed according to a preset time window; The segment log data corresponding to the preset time period before other preset time windows except the first preset time window overlap with the segment log data corresponding to the preset time period after the previous preset time window.

4. The method according to claim 1, characterized in that: The step of inputting the log data to be diagnosed into a plurality of error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model includes: For any error log diagnosis model, obtaining a fault diagnosis template of the error log diagnosis model; Inputting the to-be-diagnosed log data into the fault diagnosis template to obtain an instantiated fault diagnosis template; The instantiated fault diagnosis template is input into the corresponding error log diagnosis model to obtain the error log diagnosis result output by the error log diagnosis model.

5. The method according to claim 1, characterized in that The step of inputting the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model includes: For any error log diagnosis model, obtaining a diagnosis result voting template of the error log diagnosis model; Inputting the to-be-diagnosed log data and the error log diagnosis results output by each error log diagnosis model into the diagnosis result voting template to obtain an instantiated diagnosis result voting template; Inputting the instantiated diagnosis result voting template into the corresponding error log diagnosis model to obtain a plurality of error log diagnosis results output by the error log diagnosis model and sorted in descending order according to correctness; The voting results of the plurality of error log diagnosis results include a plurality of error log diagnosis results sorted from high to low in terms of correctness.

6. The method according to claim 1, characterized in that The voting results of the multiple error log diagnosis results include multiple error log diagnosis results sorted from high to low in terms of correctness, and the voting results of the multiple error log diagnosis results output by each error log diagnosis model are aggregated to obtain the accuracy of each error log diagnosis result, including: Based on the multiple error log diagnosis results output by each error log diagnosis model and sorted from high to low in terms of correctness, determine the correctness scores of the multiple error log diagnosis results output by each error log diagnosis model, wherein an error log diagnosis result with high correctness has a high corresponding correctness score; For any error log diagnosis model, based on the historical diagnosis accuracy of the error log diagnosis model, determine the weight of the error log diagnosis result output by the error log diagnosis model, wherein the higher the historical diagnosis accuracy, the greater the weight; For any error log diagnosis result, obtain the correctness score and weight corresponding to the error log diagnosis result under each error log diagnosis model; For any error log diagnosis model, based on the weight corresponding to the error log diagnosis result under the error log diagnosis model, weighted processing is performed on the correctness score corresponding to the error log diagnosis result under the error log diagnosis model to obtain the target correctness score corresponding to the error log diagnosis result under the error log diagnosis model; Summing target correctness scores corresponding to the error log diagnosis results under multiple error log diagnosis models to obtain the accuracy of the error log diagnosis results; Among them, the sum of the weights of the error log diagnosis results output by multiple error log diagnosis models is a preset value.

7. The method according to claim 1, characterized in that The method of determining the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result includes: Based on the accuracy of each error log diagnosis result, filter out the error log diagnosis result with the highest accuracy; Determine whether the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold; When the accuracy of the error log diagnosis result with the highest accuracy is greater than a preset accuracy threshold, an alarm is issued, and the error log diagnosis result with the highest accuracy is fed back as a fault diagnosis result of the cloud platform; When the accuracy of the error log diagnosis result with the highest accuracy is not greater than the preset accuracy threshold, the error log diagnosis result with the highest accuracy is fed back to the maintenance personnel, so that the maintenance personnel can determine the fault diagnosis result of the cloud platform based on the error log diagnosis result with the highest accuracy.

8. A cloud platform fault diagnosis device, characterized in that: The device comprises: The first acquisition module is used to obtain the error log data of the cloud platform in real time; A division module, used for dividing the error log data into a plurality of log data to be diagnosed; A second acquisition module is used for inputting any log data to be diagnosed into multiple error log diagnosis models to obtain an error log diagnosis result output by each error log diagnosis model; A third acquisition module is used to input the log data to be diagnosed and the error log diagnosis results output by each error log diagnosis model into multiple error log diagnosis models to obtain voting results of the multiple error log diagnosis results output by each error log diagnosis model; A fourth acquisition module is used to summarize the error log diagnosis results output by multiple error log diagnosis models based on the voting results of multiple error log diagnosis results output by each error log diagnosis model to obtain the accuracy of each error log diagnosis result; The determination module is used to determine the fault diagnosis result of the cloud platform based on the accuracy of each error log diagnosis result.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the cloud platform fault diagnosis method according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the cloud platform fault diagnosis method according to any one of claims 1 to 7.

Images