Search system early warning method and device based on large model and storage medium

Abstract

The invention provides a search system early warning method and device based on a large model and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of data encryption, deep learning, large models and the like. According to the specific implementation scheme, the method comprises the following steps: receiving an early warning problem from a message transmission system, wherein the early warning problem comprises a monitoring index of a search system; inputting the early warning question into an early warning large model to generate an early warning answer; and matching a corresponding early warning decision according to the early warning answer. According to the method and the device, the running state of the search system can be analyzed in real time through the large model according to the monitoring indexes of the search system, the early warning of the search system can be quickly obtained, the corresponding solution is matched, the problem can be timely found and solved before the problem is generated, and the running robustness of the search system is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to technical fields such as data encryption, deep learning, and large models. Background Art

[0002] The search system has many modules and processes, and various operational problems are prone to occur during operation. When an operational problem occurs, the search service will be interrupted, and the problem will be reported to the relevant personnel, who will solve the corresponding problem. During this period, the search system has been in an interrupted state and cannot provide services. There is a lack of early warning for the search system, the speed of discovering early warning problems is slow, and the efficiency of solving early warning problems is low. Summary of the invention

[0003] The present invention provides a search system early warning method, device and storage medium based on a large model.

[0004] According to one aspect of the present disclosure, a search system early warning method based on a large model is provided, comprising:

[0005] Receive warning issues from the messaging system, including monitoring indicators of the search system;

[0006] Input the early warning question into the early warning model to generate an early warning answer;

[0007] According to the warning answer, the corresponding warning decision is matched.

[0008] According to another aspect of the present disclosure, a search system early warning device based on a large model is provided, comprising:

[0009] A receiving module, used for receiving early warning problems from a message delivery system, the early warning problems including monitoring indicators of a search system;

[0010] An input module is used to input the warning question into the warning model to generate a warning answer;

[0011] The matching module is used to match the corresponding warning decision according to the warning answer.

[0012] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method in the embodiments of the present disclosure.

[0016] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any method according to the embodiments of the present disclosure.

[0017] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements any method according to the embodiments of the present disclosure when executed by a processor.

[0018] According to the present disclosure, a large model can be used to analyze the operating status of the search system in real time based on the monitoring indicators of the search system, quickly obtain early warnings of the search system and match corresponding solutions, so that problems can be discovered and solved in time before they occur, thereby improving the robustness of the search system operation.

[0019] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0021] Figure 1 is a flow chart of a large model-based search system early warning method according to an embodiment of the present disclosure;

[0022] Figure 2 is a flow chart of a large model-based search system early warning method according to another embodiment of the present disclosure;

[0023] Figure 3 is a flow chart of a large model-based search system early warning method according to another embodiment of the present disclosure;

[0024] Figure 4 is a flow chart of a large model-based search system early warning method according to another embodiment of the present disclosure;

[0025] Figure 5 It is a flowchart of a warning method of an intelligent search system based on a large model according to an embodiment of the present disclosure;

[0026] Figure 6 is a structural diagram of a search system early warning device based on a large model according to an embodiment of the present disclosure;

[0027] Figure 7 is a structural diagram of a search system early warning device based on a large model according to an embodiment of the present disclosure;

[0028] Figure 8is a block diagram of an electronic device for implementing an embodiment of the present disclosure. DETAILED DESCRIPTION

[0029] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0030] Figure 1 FIG. 1 is a flow chart of a large model-based search system early warning method 100 according to an embodiment of the present disclosure. In one embodiment, the method includes:

[0031] S101, receiving an early warning problem from a message delivery system, where the early warning problem includes monitoring indicators of a search system;

[0032] S102, input the early warning question into the early warning model to generate an early warning answer;

[0033] S103. Match the corresponding warning decision according to the warning answer.

[0034] In the disclosed embodiment, the search system includes multiple components or parts, which can be deployed on a server or a server cluster. The monitoring indicators of the search system may include central processing unit (CPU) monitoring indicators, memory (mem) monitoring indicators, disk (disk) monitoring indicators, queries per second (QPS) monitoring indicators, latency (latency) monitoring indicators, etc. The messaging system may include systems such as Kafka. In one example, kafka can be used to send various monitoring indicators of the search system on the server to an early warning system including multiple multimodal large models. The prompt word system of the early warning system can generate early warning questions according to the prompt word template (Prompt) based on the various monitoring indicators of the search system.

[0035] In the disclosed embodiment, the early warning problem may include monitoring indicators, and the monitoring indicators are related to the business type to which it belongs. Different business types may include different monitoring indicators. For example, for business-to-business (B2B) business, its monitoring indicators may include 97th percentile time consumption, 99th percentile time consumption, retry rate, correctness of specific resource numbers, etc. For another example, for the business of the intelligent body, its monitoring indicators may include void rate, database construction delay, average size of recall data, summary service level agreement (Service Level Agreement, SLA), estimated SLA, positive SLA, etc. The early warning problem may also carry information such as CPU usage, mem usage, disk usage, etc. within a period of time. The early warning problem is input into the trained early warning big model, and the early warning big model can analyze the early warning problem and generate an early warning answer. The early warning answer may include at least one of system error information, system error code, system alarm information, etc. The early warning big model can be fine-tuned based on the big language model, etc. using the early warning training data. The early warning training data may include information on server usage and annotated early warning answers, etc.

[0036] In the disclosed embodiment, the early warning answer can be represented in a variety of ways. For example, the early warning answer can be an early warning ID composed of a string of characters and numbers. According to the early warning answer, the corresponding early warning decision can be queried in the early warning decision system, such as the early warning decision corresponding to the early warning ID. The early warning decision may include the actions that need to be performed to solve the current early warning answer, the code that needs to be executed, the link that needs to be accessed, the program name that needs to be called, etc. For example, matching the code segment to solve the system problem corresponding to the early warning answer, reporting the system problem corresponding to the early warning answer to relevant personnel, executing the repair program, accessing the help link, etc. The early warning decision system can store the early warning answers and the early warning strategies for the system problems corresponding to the early warning answers. After obtaining the new early warning strategy, the early warning decision system can be updated in real time.

[0037] According to the embodiments of the present disclosure, the monitoring indicators of the search system can be processed through a large model, the operating status of the search system can be analyzed, the early warning answers of the search system can be quickly obtained and matched with corresponding solutions, problems can be discovered and solved in time before they occur, and the robustness of the search system operation can be improved.

[0038] In one implementation, the monitoring indicators of the search system are acquired by the message delivery system in batches from a monitoring system, and the monitoring system is used to receive the monitoring indicators uploaded by the search system.

[0039] In the disclosed embodiment, the monitoring system can be connected to each server of the search system. The monitoring indicators sent by each server are received. The server can actively report the monitoring indicators to the monitoring system at a fixed time or periodically, or report the monitoring indicators according to the instructions of the monitoring system. The monitoring system can send the monitoring indicators within a period of time to the messaging system, and send them to the large model for processing through the messaging system.

[0040] According to the embodiments of the present disclosure, the message transmission system can cache the monitoring indicators of the search system and send them to the large model for processing in batches, which can reduce the processing requirements of the large model, facilitate the use of multiple large models for parallel processing, and improve the efficiency of troubleshooting problems in the search system.

[0041] Figure 2 2 is a flow chart of a method 200 for early warning of a search system based on a large model according to another embodiment of the present disclosure. The method 200 can be used to implement step S102 in the method 100 for early warning of a search system based on a large model. In one embodiment, the method 200 includes: inputting the early warning question into the early warning large model to generate an early warning answer, and further includes:

[0042] S201, according to the target business type corresponding to the warning problem, matching the target warning big model corresponding to the target business type in multiple warning big models; wherein different business types correspond to different warning big models;

[0043] S202: Input the warning question into the target warning model to generate a warning answer corresponding to the target business type.

[0044] In the disclosed embodiment, multiple early warning big models can be trained. For example, the corresponding early warning big models can be trained according to different business types of the search system. For example, for the business type of the B2B search system, the early warning problem samples composed of one or more monitoring indicators such as the 97th percentile time consumption, the 99th percentile time consumption, the retry rate, and the correctness of a specific resource number can be used to train the early warning big model of the B2B business. For another example, for the business type of the intelligent body search system, the early warning problem samples composed of one or more monitoring indicators such as the void rate, the database construction delay, the average size of the recall data, the summary SLA, the estimated SLA, and the positive SLA can be used to train the early warning big model of the intelligent body business. According to the target business type of the early warning problem, the target early warning big model corresponding to the target business type can be selected from the big model pool. The early warning problem is input into the target early warning big model for processing, and the early warning answer corresponding to the target business type under the current early warning problem is obtained. For example, if the target business type of the early warning problem is CPU usage, according to the business type of CPU usage, the CPU usage early warning big model is matched. The CPU usage warning problem is input into the CPU usage warning big model for processing, and the CPU usage warning answer is obtained. For another example, if the target business type of the warning problem is the intelligent body business type, the intelligent body search system warning problem is input into the warning big model of the intelligent body business for processing, and the warning answer of the intelligent body business is obtained. According to the embodiment of the present disclosure, different warning big models can be used for different businesses of the search system to perform warning analysis and give corresponding warnings. Different warning big models are trained for specific business types respectively, which improves the accuracy of the warning.

[0045] In one implementation, the target service type is determined based on a target service identifier, and the target service identifier includes a message delivery address set by the message delivery system for a monitoring indicator of the target service type.

[0046] In the disclosed embodiment, after the message delivery system receives the monitoring indicators of the monitoring system, different message delivery addresses can be set for the monitoring indicators according to different business types, and the message delivery addresses can be sent to the early warning system together with the monitoring indicators as the target business identifier. For example, the message delivery system may include kafka, which is a distributed, publish / subscribe-based message system. For kafka, there are two basic types of clients: producers and consumers. Producers (also called publishers) create messages, while consumers (also called subscribers) are responsible for consuming or reading messages. In one example, the monitoring system can be used as a producer and the early warning system as a consumer. In kafka, messages are classified by topic, and each topic corresponds to a message queue. In one example, the relevant monitoring indicators of the CPU can be used as a topic, and the relevant monitoring indicators of mem can be used as another topic. In another example, the monitoring indicators of the b2b search system can be used as a topic, and the monitoring indicators of the intelligent agent search system can be used as another topic. Each topic can include multiple partitions, and the partitions can represent different message delivery channels under the topic. Each partition can send multiple messages in sequence, and each message can have a unique sequence number (offset) as the unique identifier of the message. In one example, the Topic, Partition, and offset of each monitoring indicator can be used as the target business identifier of the monitoring indicator.

[0047] According to the embodiments of the present disclosure, different monitoring indicators can be labeled according to business types, which makes it easier for the early warning system to distinguish business types. Monitoring indicators of different business types can be quickly sent to the corresponding early warning model for processing, thereby improving the speed of early warning.

[0048] Figure 3 It is a flowchart of a large model-based search system early warning method 300 according to another embodiment of the present disclosure. In one implementation, the early warning answer includes an early warning code; the method 300 can be used to implement the above-mentioned large model-based search system early warning method 100, and the method 300 also includes: S301, saving the corresponding relationship between the early warning answer and the early warning decision;

[0049] In one embodiment, the large model-based search system early warning method 300 is also used to implement step S103 in the large model-based search system early warning method 100. In one embodiment, the method 300 includes: matching the corresponding early warning decision according to the early warning answer, and further includes: according to the early warning code, querying the early warning decision corresponding to the early warning code in the corresponding relationship.

[0050] In the disclosed embodiment, the correspondence between the warning answer and the warning decision can be stored in the warning decision system. After obtaining the warning answer, the corresponding relationship can be searched according to the warning answer to obtain the corresponding warning decision. In one example, the correspondence between the warning answer and the warning decision can be set in advance based on experience and warning processing history, and the warning answer, warning decision and the corresponding relationship can be stored in the warning decision system. In another example, after obtaining the warning decision corresponding to the new warning answer, the new warning answer, warning decision and the corresponding relationship can be stored in the warning decision system.

[0051] In the embodiments of the present disclosure, the warning answer may include a warning code, and different warning codes may correspond to different specific warning problems. For example, the warning code "1" may correspond to the warning problem "CPU occupancy is too high", etc., and the warning code "2" may correspond to the warning problem "mem occupancy is too high", etc., which is not limited in the present disclosure. In the warning decision system, the warning code and the warning decision to solve the warning problem corresponding to the warning code may be stored according to the corresponding relationship. After obtaining the warning answer using the large model, the corresponding warning decision can be found according to the warning code of the warning answer.

[0052] According to the embodiments of the present disclosure, a corresponding warning decision can be obtained by querying in the warning decision system based on the warning answer, thereby improving the speed of solving the warning and improving the efficiency of solving the warning problem.

[0053] Figure 4 4 is a flow chart of a method 400 for early warning of a search system based on a large model according to another embodiment of the present disclosure. The method 400 can be used to implement S103 in the method 100 for early warning of a search system based on a large model. In one embodiment, the method 400 includes: matching a corresponding early warning decision according to the early warning answer, and further includes at least one of the following:

[0054] S401, executing the code included in the early warning decision to solve the early warning problem;

[0055] S402: Send an early warning prompt to the receiving end according to the early warning decision.

[0056] In the disclosed embodiment, the early warning decision may include sending early warning prompts and executing codes to solve early warnings. In the execution code to solve the early warning method, the code corresponding to the early warning answer can be directly set, and the code can be executed to solve the early warning problem. In the sending early warning prompt method, according to the business type of the early warning answer, it can be sent to the corresponding client, and the relevant personnel responsible for the business can solve the early warning problem.

[0057] According to the embodiments of the present disclosure, when the warning can be directly resolved by executing the code, the code is executed to solve the problem, thereby improving the efficiency of resolving the warning. New warnings or warnings that lack corresponding codes can be reported in a timely manner to relevant personnel of the business to which the warning belongs, thereby improving the efficiency of warning reporting, reducing the time for discovering and handling problems, and maintaining the operating status of the search system.

[0058] In one embodiment, the early warning big model includes an offline early warning big model and an online early warning big model; the training data of the offline early warning big model includes question and answer prompt words, and the question and answer prompt words include the monitoring indicators of the search system and their corresponding early warning answers; the training data of the online early warning big model includes question and answer prompt words that meet the training objectives of the offline early warning big model.

[0059] In the disclosed embodiment, the initial offline warning big model can be the same as or similar to the online warning big model. First, use the offline warning big model to obtain question prompt words that meet the training objectives. Then train the online warning big model based on the question and answer prompt words of the training objectives. Specifically, after entering the question and answer prompt words in the offline warning big model, the predicted answer can be output by processing the questions therein. The offline warning big model can be fine-tuned based on the output answer and the answer in the question and answer prompt words. If the initial offline warning big model can be adjusted to meet the training objectives based on one or more question and answer prompt words. These one or more question and answer prompt words can be used as training data for the online warning big model.

[0060] In the disclosed embodiment, the question and answer prompt words in the training data may include monitoring indicators of the search system collected in advance and their corresponding existing warning answers, which can be obtained based on past experience and problem solving history.

[0061] According to the embodiments of the present disclosure, the offline early warning big model can be used for simulation training to filter out training data for the online early warning big model. The use of these training data can more quickly and accurately adjust the online early warning big model, so that the online early warning big model can give more accurate early warning answers for real monitoring indicators, improve the authenticity of the early warning big model's prediction results, and increase the applicability of the early warning big model to real situations.

[0062] In one implementation, when the training data of the offline early warning model does not meet the training objectives of the offline early warning model, the training data of the offline early warning model includes updated question and answer prompt words.

[0063] In the disclosed embodiment, if the initial offline warning model cannot be adjusted to meet the training target based on one or more question and answer prompt words, the adjusted offline warning model can be reset to the initial offline warning model. Then a new batch of question and answer prompt words are used to train the reset offline warning model. The above process is repeated until the output result of the offline warning model for the adjusted question and answer prompt words meets the training target, and then the question and answer prompt words that meet the training target can be used as training data for the online warning model.

[0064] According to the embodiments of the present disclosure, training data that meets the training objectives can be provided for the online early warning big model, thereby improving the training effect and training efficiency of the online early warning big model, and obtaining an online early warning big model that can output more accurate early warning answers based on real monitoring indicators.

[0065] The present disclosure proposes a method for early warning of an intelligent search system based on a large model, which can detect problems in time before they occur and report them to relevant personnel. The process of the method for early warning of an intelligent search system based on a large model is as follows: Figure 5 As shown. This method can be used to timely warn relevant personnel before search problems occur, reducing losses and shortening repair time. The method includes: a large model system capable of generating early warning information; a prompt word generation system based on early warning data; and an early warning decision system.

[0066] The large model system of early warning information can receive relevant early warning information. Relevant early warning information may include processor (CPU) usage, memory (mem) usage, disk (disk) usage, etc. within a period of time, and output a warning answer whether the system has a problem. Input the standard question and answer (Q&A) prompt words into the offline early warning large model, and when the output answer meets the requirements, use the prompt words to fine-tune the online early warning large model. When the output answer does not meet the requirements, modify the standard Q&A prompt words, reset the offline early warning model, and execute this step again.

[0067] The prompt word system can generate standard Q&A prompt words based on training samples. The training samples include pre-collected search server monitoring indicators (cases) and corresponding warning answers. Based on the standard Q&A prompt words, the warning large model can be fine-tuned on the basis of the original large model. Each case includes information such as CPU, mem, disk, QPS, latency, etc., and the corresponding warning answer is the preset warning answer corresponding to the case. The prompt word system receives the search server monitoring indicators sent by the upstream messaging system (kafka), and composes the search server monitoring indicators into warning questions. The search server monitoring indicators are the various indicators of the server where the search engine is located that are written into the messaging system by the monitoring system.

[0068] The prompt word generation system can also receive relevant monitoring indicators sent by the upstream information transmission system (kafka), and then compose these indicators into questions to ask the big model, and get the early warning answer. After obtaining the monitoring indicators, the monitoring system can write the indicators into kafka for transmission to the big model. The messaging system can set different identification information (such as kafka address) for different business types. After sending it to the early warning system, the early warning system distributes it to different early warning big models for processing according to the identification information of the received early warning questions. The early warning big model generates early warning answers based on the early warning questions. The early warning questions reflect the CPU usage, mem usage, disk usage, etc. over a period of time. The early warning answers include error codes.

[0069] The early warning decision system can determine the early warning decision based on the error code of the early warning answer and the pre-set program. The early warning decision system provides an interface. After the prompt word generation system gets the answer, it can request the early warning decision system based on the answer. The system will decide whether to notify relevant personnel to investigate based on the answer and the pre-set program.

[0070] The disclosed embodiments can promptly warn users before search system problems occur, thereby reducing repair time and minimizing losses.

[0071] Figure 6 6 is a schematic diagram of a structure of a search system early warning device 600 based on a large model according to an embodiment of the present disclosure. In one embodiment, the device includes:

[0072] A receiving module 601 is used to receive a warning problem from a message delivery system, where the warning problem includes a monitoring indicator of a search system;

[0073] Input module 602, used to input the warning question into the warning big model to generate a warning answer;

[0074] The matching module 603 is used to match the corresponding warning decision according to the warning answer.

[0075] In one implementation, the monitoring indicators of the search system are acquired by the message delivery system in batches from a monitoring system, and the monitoring system is used to receive the monitoring indicators uploaded by the search system.

[0076] Figure 7 7 is a schematic diagram of a structure of a large model-based search system early warning device 700 according to an embodiment of the present disclosure. The device 700 includes: a receiving module 701, an input module 702 and a matching module 703. The functions of the above modules can refer to the functions of the modules of the re-illumination device 600 in the above embodiment. In one embodiment, the input module 702 includes:

[0077] The matching submodule 7021 is used to match the target warning big model corresponding to the target business type in multiple warning big models according to the target business type corresponding to the warning problem; wherein different business types correspond to different warning big models;

[0078] The generation submodule 7022 is used to input the warning question into the target warning model to generate a warning answer corresponding to the target business type.

[0079] In one implementation, the target service type is determined based on a target service identifier, and the target service identifier includes a message delivery address set by the message delivery system for a monitoring indicator of the target service type.

[0080] In one embodiment, the warning answer includes a warning code; Figure 7 As shown, the device 700 further includes: a storage module 704, which is used to store the corresponding relationship between the warning answer and the warning decision;

[0081] In one implementation, the matching module 703 is further configured to query the warning decision corresponding to the warning code in the corresponding relationship according to the warning code.

[0082] In one embodiment, Figure 7 As shown, the matching module 703 further includes at least one of the following:

[0083] An execution submodule 7031 is used to execute the code included in the warning decision to solve the warning problem;

[0084] The sending submodule 7032 is used to send a warning prompt to the receiving end according to the warning decision.

[0085] In one embodiment, the early warning big model includes an offline early warning big model and an online early warning big model; the training data of the offline early warning big model includes question and answer prompt words, and the question and answer prompt words include the monitoring indicators of the search system and their corresponding early warning answers; the training data of the online early warning big model includes question and answer prompt words that meet the training objectives of the offline early warning big model.

[0086] In one implementation, when the training data of the offline early warning model does not meet the training objectives of the offline early warning model, the training data of the offline early warning model includes updated question and answer prompt words.

[0087] For the description of specific functions and examples of each module and submodule of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0088] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0089] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0090] Figure 8 A schematic block diagram of an example electronic device 800 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0091] like Figure 8 As shown, the device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0092] A number of components in the device 800 are connected to the I / O interface 805, including: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a disk, an optical disk, etc.; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0093] The computing unit 801 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 801 performs the various methods and processes described above, such as a search system early warning method based on a large model. For example, in some embodiments, the search system early warning method based on a large model may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded into the RAM 803 and executed by the computing unit 801, one or more steps of the search system early warning method based on the large model described above may be executed. Alternatively, in other embodiments, the computing unit 801 may be configured to execute the large model-based search system early warning method in any other appropriate manner (eg, by means of firmware).

[0094] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0095] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0096] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0097] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0098] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0099] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0100] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0101] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A search system early warning method based on a large model, comprising: Receiving an early warning problem from a messaging system, the early warning problem including monitoring indicators of a search system; Input the early warning question into the early warning model to generate an early warning answer; According to the warning answer, the corresponding warning decision is matched.

2. The method according to claim 1, wherein: The monitoring indicators of the search system are acquired by the message transmission system from the monitoring system in batches, and the monitoring system is used to receive the monitoring indicators uploaded by the search system.

3. The method according to claim 1, wherein: Input the early warning question into the early warning model to generate early warning answers, including: According to the target business type corresponding to the warning problem, a target warning big model corresponding to the target business type is matched in multiple warning big models; wherein different business types correspond to different warning big models; The warning question is input into the target warning big model to generate a warning answer corresponding to the target business type.

4. The method according to claim 3, wherein: The target service type is determined based on a target service identifier, and the target service identifier includes a message delivery address set by the message delivery system for a monitoring indicator of the target service type.

5. The method according to claim 1 or 2, wherein: The early warning answer includes an early warning code; The method further comprises: saving the corresponding relationship between the warning answer and the warning decision; According to the warning answer, the corresponding warning decision is matched, including: according to the warning code, searching the corresponding relationship for the warning decision corresponding to the warning code.

6. The method according to any one of claims 1 to 5, wherein: According to the warning answer, matching the corresponding warning decision also includes at least one of the following: Execute the code included in the early warning decision to solve the early warning problem; Send an early warning prompt to the receiving end according to the early warning decision.

7. The method according to claim 1, wherein: The early warning big model includes an offline early warning big model and an online early warning big model; the training data of the offline early warning big model includes question and answer prompt words, and the question and answer prompt words include the monitoring indicators of the search system and their corresponding early warning answers; the training data of the online early warning big model includes question and answer prompt words that meet the training objectives of the offline early warning big model.

8. The method according to claim 7, wherein: When the training data of the offline early warning model does not meet the training target of the offline early warning model, the training data of the offline early warning model includes updated question and answer prompt words.

9. A search system early warning device based on a large model, comprising: A receiving module, used to receive early warning issues from a message delivery system, wherein the early warning issues include monitoring indicators of a search system; An input module, used to input the warning question into the warning big model to generate a warning answer; A matching module is used to match the corresponding warning decision according to the warning answer.

10. The device according to claim 9, wherein: The monitoring indicators of the search system are acquired by the message transmission system from the monitoring system in batches, and the monitoring system is used to receive the monitoring indicators uploaded by the search system.

11. The device according to claim 9, wherein: The input module comprises: A matching submodule, for matching a target early warning big model corresponding to the target business type in a plurality of early warning big models according to the target business type corresponding to the early warning problem; wherein different business types correspond to different early warning big models; The generation submodule is used to input the warning question into the target warning model to generate a warning answer corresponding to the target business type.

12. The device according to claim 11, wherein: The target service type is determined based on a target service identifier, and the target service identifier includes a message delivery address set by the message delivery system for a monitoring indicator of the target service type.

13. The device according to claim 9 or 10, wherein: The warning answer includes a warning code; the device also includes: a storage module, which is used to store the corresponding relationship between the warning answer and the warning decision; The matching module is further used to query the warning decision corresponding to the warning code in the corresponding relationship according to the warning code.

14. The device according to any one of claims 9 to 13, wherein: The matching module further includes at least one of the following: An execution submodule, used for executing the code included in the early warning decision to solve the early warning problem; The sending submodule is used to send a warning prompt to a receiving end according to the warning decision.

15. The device according to claim 9, wherein: The early warning big model includes an offline early warning big model and an online early warning big model; the training data of the offline early warning big model includes question and answer prompt words, and the question and answer prompt words include the monitoring indicators of the search system and their corresponding early warning answers; the training data of the online early warning big model includes question and answer prompt words that meet the training objectives of the offline early warning big model.

16. The device according to claim 15, wherein: When the training data of the offline early warning model does not meet the training target of the offline early warning model, the training data of the offline early warning model includes updated question and answer prompt words.

17. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-8.

19. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

20. A large model comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images