Server system, information processing device, control method for information processing device, and program

Abstract

To clearly inform the user of the nature of the error based on the error log. [Solution] A server system is provided that exchanges information between an application server that runs an application, an LLM server having a large-scale language model, and a notification server that provides notification services, the server system comprising: a detection unit that detects errors that occur in the application server; a first acquisition unit that acquires an error log corresponding to the error detected by the detection unit and a prompt; a second acquisition unit that acquires analysis information showing the result of analyzing the error log in the LLM server based on the prompt; and a notification unit that controls the notification of the analysis information acquired by the second acquisition unit to the user via the notification server.

Description

Technical Field

[0001] The present disclosure relates to a server system, an information processing apparatus, a control method for an information processing apparatus, and a program.

Background Art

[0002] In an information processing apparatus such as a computer, when an abnormality occurs in the system, a technique for notifying the user of the abnormality is known. In Patent Document 1, when an abnormality occurs in the system, based on allocation data that associates an error ID for identifying the abnormality with information on the software configuration, a technique for displaying an alarm on the screen regarding the function of the software corresponding to the abnormality is disclosed.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] There is a need for a technique for notifying the user of the content of the error in an easy-to-understand manner based on the error log.

Means for Solving the Problems

[0005] A server system according to one aspect of the present disclosure is a server system that exchanges information between an application server that executes an application, an LLM server having a large-scale language model, and a notification server that provides notification services, and is characterized by comprising: a detection unit that detects errors occurring in the application server; a first acquisition unit that acquires an error log corresponding to the error detected by the detection unit and a prompt; a second acquisition unit that acquires analysis information showing the result of analyzing the error log in the LLM server based on the prompt; and a notification unit that controls the notification of the analysis information acquired by the second acquisition unit to a user via the notification server. [Effects of the Invention]

[0006] According to this disclosure, the content of the error can be clearly communicated to the user based on the error log. [Brief explanation of the drawing]

[0007] [Figure 1] This is a diagram showing an example of the configuration of a notification system. [Figure 2] This is a block diagram showing an example of the hardware configuration of an information processing device. [Figure 3] This figure shows an example of the configuration of a user terminal. [Figure 4] This figure shows an example of a platform server configuration. [Figure 5] This figure shows an example of the UI screen of an expense claim application. [Figure 6] This figure shows an example of the UI screen of a prompt creation application. [Figure 7] This figure shows an example of a notification server configuration. [Figure 8] This figure shows an example of the UI screen of a notification application. [Figure 9] This sequence diagram shows an example of the process from error log analysis to notification of the results. [Figure 10]This flowchart shows an example of the processes performed by the platform server. [Figure 11] This flowchart shows an example of the process performed by the notification server. [Figure 12] This figure shows an example of a platform server configuration. [Figure 13] This flowchart shows an example of the processes performed by the platform server. [Figure 14] This figure shows an example of a notification server configuration. [Figure 15] This flowchart shows an example of the process performed by the notification server. [Modes for carrying out the invention]

[0008] Preferred embodiments of the server system, information processing device, control method for the information processing device, and program relating to this disclosure will be described in detail below with reference to the drawings. However, the following embodiments do not limit the server system, information processing device, control method for the information processing device, and program relating to this disclosure. Furthermore, identical components are denoted by the same reference numerals in the following embodiments, and redundant descriptions are omitted.

[0009] In information processing devices such as computers, system malfunctions can occur. Furthermore, not limited to system malfunctions, errors can occur due to user actions while a user is using software or multiple applications running on a platform. In such cases, technologies are desired to quickly recover from system malfunctions or resolve the causes of software errors. One such technology is known to display an alarm on the screen regarding the software function corresponding to a system malfunction, based on assignment data that links an error ID to software configuration information. However, system malfunctions and user-initiated errors are numerous and may not always be uniquely matched to the software configuration. Also, even if information indicating the malfunction or error is notified to the user, the user may not immediately understand its meaning.

[0010] The embodiments described below illustrate an example in which the user can immediately understand the nature of a system anomaly or error.

[0011] In this embodiment, a large language model (LLM) capable of outputting text data and other information according to a prompt is used, based on the input of a prompt that provides instructions regarding the output result. For example, the user can provide detailed settings and instructions in a prompt, and then input the necessary data along with this prompt into the large language model, thereby obtaining the analysis results of that data, which has been analyzed according to the settings and instructions, as the output result. In this embodiment, if a system anomaly occurs or an error occurs due to user operation, a pre-prepared prompt and information indicating the anomaly or error are input into the large language model. The output result then notifies the user of the anomaly or error in an easy-to-understand manner. This allows the user to immediately understand the nature of the system anomaly or error.

[0012] <<First Embodiment>> FIG. 1 is a diagram showing an example of the configuration of the notification system in the present disclosure. This notification system is a server system including a user terminal 101, a platform server 102, an LLM server 103, and a notification server 104. These are communicably connected to each other via a network 105 and exchange information (such as data) as described later. The notification system shown in FIG. 1 includes one user terminal 101, one platform server 102, one LLM server 103, and one notification server 104 each, but may include two or more of each device. For example, this notification system may include a server group functioning as the platform server 102, a server group functioning as the LLM server 103, and a server group functioning as the notification server 104. Alternatively, this notification system may be one server having the functions of the platform server 102, the LLM server 103, and the notification server 104.

[0013] (Regarding the user terminal 101) The user terminal 101 is a terminal device possessed by a user who uses services provided by the platform server 102 and notification services provided by the notification server 104, which will be described later. Note that the user terminal 101 may also be capable of using services provided by the LLM server 103, which will be described later.

[0014] The user terminal 101 will be described by taking, for example, a Personal Computer (PC) as an example. Note that the user terminal 101 may be a desktop PC, a notebook PC, a tablet terminal, or a smartphone, etc. That is, the user terminal 101 may be any information processing device capable of executing the functions or applications described later.

[0015] User terminal 101 has an application (hereinafter referred to as the notification application) installed for using notification services provided by notification server 104. By running the notification application, user terminal 101 can provide notification services to users using user terminal 101. In addition, user terminal 101 has a general browser installed. By using the browser, users can access the server indicated by the specified Uniform Resource Locator (URL) and view web pages, etc.

[0016] Furthermore, by accessing the platform server 102 from the user terminal 101, the user can receive various services. In this disclosure, an example of using a browser function as a method of accessing the platform server 102 will be described. For example, the user enters a URL indicating the platform server 102 into an input area provided on the browser screen displayed on the display unit of the user terminal 101. Then, by entering a user ID and password on the authentication screen displayed on the browser, the user can access the platform server 102 via the user terminal 101 and use various services and functions.

[0017] (Regarding platform server 102) Platform server 102 is an application platform server that, through its functions, runs multiple applications and provides the functions and services of these applications. In this disclosure, the platform server 102 will be described as providing services related to the application and approval procedures for expenses in a company, and services related to the creation of prompts to be sent to LLM server 103. In addition, platform server 102 may provide various other services. For example, platform server 102 can run applications developed by the administrator of platform server 102 using its functions, thereby providing the functions and services of those applications to users. Furthermore, the functions and services of these applications may be provided by separate platform servers. That is, platform server 102 may be a group of application servers capable of running at least one or more applications.

[0018] Services related to expense claims and approval procedures within a company can be implemented by an application that provides such services (hereinafter referred to as the expense claim application). When the expense claim application is executed on the platform server 102, users can enter strings or other information into input fields corresponding to items for expense claims, or select options corresponding to those items, via the user terminal 101. The expense claim application can determine whether the necessary information for the expense claim has been entered and selected, and can notify the approver of the claim that the claim has been submitted.

[0019] The service related to prompts sent to the LLM server 103 can be implemented by an application that provides that service (hereinafter referred to as the prompt creation application). By running the prompt creation application on the platform server 102, the user can create prompts to send to the LLM server 103 via the user terminal 101. The user can also simulate the response from the LLM server 103 to those prompts.

[0020] (Regarding LLM Server 103) The LLM server 103 is a server that possesses and provides the functionality of a large-scale language model. The large-scale language model can output output results according to instructions indicated by input prompts. In addition, when the large-scale language model receives arbitrary data, it can analyze that data according to instructions indicated by prompts and output the results of that analysis.

[0021] The LLM server 103 can provide an Application Programming Interface (API) for utilizing this function. Furthermore, if API integration with other devices such as the platform server 102 is possible, ChatGPT®, Claude®, or Gemini® may be used as the large-scale language model in this disclosure. Since publicly known technologies can be used for the format of requests and responses using the API, a detailed explanation is omitted.

[0022] When the LLM server 103 receives prompts and error logs, for example via an API, it inputs them into the large-scale language model and sends the output results back to the source of the prompts and error logs.

[0023] <Example of a prompt> The following is an example of a prompt sent from the platform server 102 to the LLM server 103. Note that the following prompt can be created by the user using a prompt creation application, as described later.

[0024] ----------------------------------------------------------------------------- You are a **"YYYY Specialist AI Monitor"**. Analyze the given error log and output a summary and evaluation based on the following rules. ### Evaluation Rules (Applied in order of priority) 1. **Whitelist Check** - Error message containing "role required" → :information_source: Information -Error message contains "Registered role required" →:information_source:information - The error message includes "If SelectableOption is FALSE, the workflow code must be "void"" → :information_source: Information - The error message includes "Items that violate the job authority regulations cannot be set" → :information_source: Information -Error message contains "No XXXXUser found to check role" →:information_source:information - The error message contains "Audit_Events: String_of_BuildVer did not find attribute for record Workflow_Approval" → :information_source: information 2.**Check specific rules** -Module is "SQL" → :information_source: information -Module is "ServiceCenter" → :warning:Caution (Possible management system failure) -Module is "Timer" → Evaluated by delay time - "took more than N seconds," etc. → :warning:Caution -Failure and not re-executed → :rotating_light: Emergency -Module is "Integration" → External system integration error (depending on the content: :warning:Caution or :rotating_light:Emergency) 3. **Special Rules** -If "Invalid username or password" occurs: - Occurrences 5 times or less →:information_source: Information -6 or more occurrences (repeated occurrences) → :warning:Caution (Possible fraudulent login attempt or configuration error) 4. **Evaluation based on general standards (other than those mentioned above)** -:rotating_light:Critical: Persistent error or business disruption level (e.g., "Object reference...", "Violation of PRIMARY KEY...", "Process ... has been suspended") -:warning:Caution: Repeated performance degradation or partial malfunction (e.g., "Timeout" repeatedly occurring on the same screen) -:information_source:Info:Temporary errors that can be automatically recovered (e.g., one-time "Timeout expired", one-time "Invalid username or password") 5. **Promotion Rules Based on Frequency and Impact** - Single occurrence → Information (Info) - Repeated occurrences with the same function → Warning - Overall impact or permanent occurrence → Critical ### Output Format Please output your Markdown file in the following order and labels. [Urgency]: {Show urgency} [Tenant Information]: {Display Tenant (If Tenant=Specific_Items, convert to XXXX) (If Tenant=38, convert to XXXX)} [Time of occurrence / Error ID]: {If ID and Time of Log are available → "yyyy-mm-dd hh:mm:ss" / xxx} {If only an ID is available, use "xxx"} {If only Time is available → "yyyy-mm-dd hh:mm:ss"} [Error Summary]: {Summarize the core of the error in 1-2 sentences} [Estimated scope of impact]: {Estimate what is happening in which function. If possible, specify the screen name or function name.} -If the Message contains the Entity name / Action name, please specify it. -If only the module name is provided, it will be labeled as "Processing within the relevant module". - No additional information available → "Cannot be confirmed" is displayed. [Recommended initial actions]: {Specific actions are provided for each error type} -DB connection error → "Check the operational status of the connection pool / DB" - Primary Key violation → "Check the app's duplicate registration logic" -Timer delay → "Check the logs of the relevant Timer and consider whether the processing time can be reduced." - External API error → "Check Integration logs and external system operational status" -Other unknowns → "Check detailed logs in Service Center" -Tenant / ID / Time of Log should only be output if they are included in the input. ### Term definition -Draft="body" -Attachment="Attachments in the main text" -Request=「Request」 -Tenant=“Department in charge” ### Restrictions - Do not answer based on speculation. If information is unknown, do not make assumptions; honestly state, "We cannot confirm this." - The default database will be POSTGRESQL (registered trademark). ### Example of execution Input (Example error): Message: "Timeout expired. The timeout period elapsed prior to obtaining a connection from the pool.", Module: SQL, Tenant: ******, Id: 12345, Time of Log: 2025-08-01 14:32:10 ------------------------------------------------------------------------------ By setting a personality and defining the direction of thought, such as "You are **'YYYY specialized AI monitor'**" at the beginning of the above prompt, it is possible to obtain highly accurate output results.

[0025] Furthermore, by instructing the system to output a summary and evaluation of the error content at the prompt, the error content shown in the error log can be output. In this disclosure, evaluation refers to the importance or urgency of the error content, with importance increasing in the order of "Informative," "Caution," and "Urgent." For example, "Informative" in evaluation indicates an importance level sufficient to notify that an error has occurred. "Caution" in evaluation indicates a higher importance level than "Informative," indicating that attention is required regarding the error that has occurred. "Urgent" in evaluation indicates a higher importance level than "Caution," indicating that immediate action is required regarding the error that has occurred.

[0026] Furthermore, by defining detailed evaluation rules, output formats, terminology, constraints, and execution examples in the prompt, consistency can be ensured in the output format for each output result. This allows users to quickly understand the content notified as an output result. The following explains the "evaluation rules" and "output format" among the items described in the prompt.

[0027] (Regarding evaluation rules) The evaluation rules are regulations for determining the severity of an error. In other words, the evaluation rules define how errors that may occur on the platform server 102 can be categorized into multiple severity levels. Here, the whitelist in the evaluation rules is a list of low-severity error messages. For example, the whitelist includes cases where the error log's "Message" indicates "role required" or "Registered role required" as the error message. Errors that occur when an application procedure that should be performed with the required privileges is attempted by a user without the necessary privileges are considered low-severity and are therefore included in the whitelist.

[0028] The specific rules in the evaluation rules are provisions for changing the severity of an error depending on what the "Module" in the error log indicates. For example, one specific rule stipulates that if "Module" indicates "ServiceCenter," there may be a problem related to the management of platform server 102, and therefore the evaluation should be changed to "Caution" and a notification should be sent.

[0029] Special rules in evaluation rules are provisions for changing the severity of errors under specific conditions. For example, a special rule might stipulate that if an inappropriate user ID or password is entered six or more times, this input may indicate unauthorized access, and the evaluation should be changed to "caution" and a notification should be issued.

[0030] In addition to the rules mentioned above, the prompt also describes the provision for evaluation based on general standards and promotion rules based on frequency and impact. Note that the evaluation rules mentioned above are not limited to these, and may include other necessary provisions, or some of the above items may be omitted.

[0031] (Regarding output format) The output format specifies the format of the output results output from LLM server 103. The output format consists of "urgency," "tenant information," "time of occurrence / error ID," "error summary," "estimated scope of impact," and "recommended initial actions."

[0032] The "urgency level" indicates the severity of the error and is shown as ":information_source:information", ":warning:warning", or ":rotating_light:urgent".

[0033] "Tenant information" indicates the recipient of error notifications. For example, if the recipient is a user using a service provided by platform server 102, "Tenant information" will display a string indicating that user.

[0034] The "Time of Occurrence / Error ID" field indicates the time the error occurred and an identifier used to identify that error. The content of the "Time of Occurrence / Error ID" field will vary depending on the contents of the error log.

[0035] The "Error Summary" section describes the nature of the error that has occurred. For example, it should concisely explain the error so that users can understand what is happening just by looking at this section. The "Estimated Impact" section indicates where the error is occurring and the extent of its impact. The "Recommended Initial Actions" section indicates the steps the user should take to address the error.

[0036] The output format configuration is not limited to the configuration described above; other configurations may be included, and some of the above configurations may be omitted. Furthermore, not only the "evaluation rules" and "output format," but also the "term definitions," "constraints," and "execution examples" can be configured by the user at the prompt. For example, by instructing the user to describe in the "constraints" section if the content or cause of an error cannot be identified or inferred, the notification of incorrect information can be reduced. In other words, instructions at the prompt can enable notification to the user even for errors that are difficult to link to error messages.

[0037] <Regarding the contents of the error log> The following is an example of the contents of an error log sent from platform server 102 to LLM server 103. The error log below is a log of an error that occurred on platform server 102 and is stored in the memory of platform server 102. Examples of errors include errors that occur when a user uses an application running on platform server 102 via user terminal 101, and errors that occur due to a system abnormality on platform server 102.

[0038] ----------------------------------------------------------------------------- {"AppName":" 215_Proxy Approver Configuration", "Environment_Information":"eSpaceVer: Id=12345, PubId=0, CompiledWith=123.4.5.6789\r\nRequestUrl: http: / / yyyy -dev.eem.xxxx-corp.co.jp / PrimaryServiceModule / screenservices / PrimaryServiceModule / MainFlow / TargetEntities / ScreenDataSetGetTargetEntities (Method: POST)\r\nAppDomain: / LM / W3SVC / 1 / ROOT / AssignedOverseerProfile-11-134029812250123456\r\nFilePath: C:\\...\\PS\\running\\AssignedOverseerProfile.12345678910 \\\r\nClientIp: 10.61.1.234 X-Forwarded-For: 111.11.111.111\r\nLocale: ja-JP\r\nDateFormat: yyyy-MM-dd\r\nPID: 1234 ('w3wp', Started='2025 / 09 / 22 2:06:37', Priv=1561Mb, Virt=2123502Mb)\r\nTID: 319\r\nThread Name: \r\n.NET: 2.0.30041.41000", "id":"12345678-90ab-cdef-1234-567890abcdef", "Message":"Registered role required", "Module":"AssignedOverseerProfile", "Server":"YYYY-FEC-DEV-A", "SessionId":"B+3UkzPayLk20Z1234Ly3e==", "Source":"AssignedOverseerProfile", "Stack":"Registered role required\r\nLocation ssAssignedOverseerProfile.ScreenServices.AssignedOverseerProfile_MainFlow_AlternateReviewers_Controller.CheckPermissions(HeContext heContext)\r\nLocation ssAssignedOverseerProfile.ScreenServices.AssignedOverseerProfile_MainFlow_AlternateReviewers_Controller. <screendatasetgetsurrogatehogetus>b__1(HeContext heContext, String screenName, JObject screenModel, JObject inputParameters, JObject clientVariables)\r\nLocation yyyy.RESTService.Controllers.ScreenServicesApiController.endpoint(String input, String endpointName, String apiVersion, EndpointImplementationDelegate implementation)", "Tenant": "22", "Time_of_log":"2025-09-22T04:04:06.88811234", "ErrorCount":3} ----------------------------------------------------------------------------- While a detailed explanation of the error log is omitted here, the error log includes information such as the date and time the error occurred, the nature of the error, the location where the error occurred, and the number of times the error occurred. Even experienced users may find it difficult to quickly understand the contents of the error log at a glance. Furthermore, while users may easily understand error logs they frequently encounter, they may find it difficult to understand error logs they are seeing for the first time.

[0039] If such error logs were directly notified to users, users would need to refer to dedicated manuals or other resources to identify the nature and location of the error, and then consider appropriate countermeasures. Furthermore, if error logs were notified to users regardless of their severity, users would be forced to follow the same procedures even for errors that do not require action, leading to inefficiency and increasing the burden on users.

[0040] Therefore, in order to alleviate such burdens on users, it is desirable to notify users of a brief explanation of the error and the corresponding measures to be taken when an error occurs on the platform server 102.

[0041] <About the output results> The following is an example of the output results (sometimes referred to as analysis information) that are output when the above-mentioned prompts and error logs are input to the large-scale language model of the LLM server 103.

[0042] ----------------------------------------------------------------------------- [Urgency]: :information_source: Information (Info) [Tenant Information]: XXXX [Time of occurrence / Error ID]: 2025-09-22 04:04:06 / 12345678-90ab-cdef-1234-567890abcdef [Error Summary]: The error message "Registered role required" occurred, indicating that a registered role is required for the permission check process on a specific delegate approver settings screen. [Estimated scope of impact]: This issue occurs in the ScreenDataSetGetTargetEntities method within the MainFlow of the AssignedOverseerProfile module, affecting the processing of the permission delegation settings screen. [Recommended initial action]: This error is an informational error related to whitelisting. Check your permission settings and consider registering any missing roles. Check the logs in Service Center for more details. ----------------------------------------------------------------------------- The output results above are presented in the format specified in the prompt above. For example, the "Error Summary" in the output results provides a brief explanation of the error and its cause. The "Estimated Impact" provides a brief explanation of the scope of the impact caused by the error. The "Recommended Initial Actions" indicates the severity of the error and the actions the user should take.

[0043] It goes without saying that the content of the output is not limited to this and will change depending on the content of the prompt and error log. Furthermore, by changing the instructions or settings in the prompt, the output may be made to show more detailed information about the error, its scope of impact, and the actions the user should take to address it.

[0044] (Regarding notification server 104) The notification server 104 is a server that provides notification services. For example, a user can use notification services by using the notification application provided by the notification server 104. As preparation for using notification services, the user launches the notification application on the user terminal 101 and registers as a user. This registration information is stored in the memory of the notification server 104. The notification server 104 assigns a user ID to the user during registration. The user can then use notification services, such as viewing notifications from other registered users, by using the notification application. In addition, the notification server 104 can provide a chat service that allows users to send notifications of text and other information to each other via the notification application, as well as a network calling service that allows calls using video and audio.

[0045] Furthermore, users who utilize these services are assigned a channel ID, which is a different identifier from their user ID, by the notification server 104. A channel represents a group of users (sometimes referred to as a group). Users can freely configure channels. For example, a user belonging to a company can configure channels based on their company's department or team. Users can also add users to a configured channel or remove users from a channel. The channel ID is an identifier used to identify a group of users. The notification server 104 can also provide services such as notification services to users via the web browser on the user terminal 101.

[0046] The notification server 104 can provide an API for using these services. For example, a user prepares a channel ID registered with the notification server 104. The platform server 102 sends the channel ID and the analysis results of the error log received from the LLM server 103 to the notification server 104's API in the API request. When the notification server 104 receives the analysis results of the error log via the API, it can notify the group indicated by that channel ID of the analysis results. Alternatively, a user ID may be used instead of a channel ID for notifications using the API. The notification of analysis results may also be implemented using technologies such as webhooks.

[0047] Furthermore, notification services, etc., are not particularly limited, but may be provided via, for example, Slack® or Teams®, as long as they can be linked via API with other devices such as the platform server 102.

[0048] (Hardware configuration of information processing equipment) The user terminal 101, platform server 102, LLM server 103, and notification server 104 described above are implemented by an information processing device 200 ("an example of a computer") with a configuration as shown in Figure 2. Figure 2 is a block diagram showing an example of the hardware configuration of the information processing device 200. The information processing device 200 has a CPU 201, ROM 202, RAM 203, storage device 204, input / output I / F 205, and network interface (I / F) 206. These components are connected to each other by a system bus. In addition, a display unit 207 and an input unit 208 are connected to the input / output I / F 205.

[0049] The CPU 201 reads programs stored in the ROM 202 and the memory device 204 into the RAM 203, and executes the read programs to perform various controls. The CPU 201 may be an MPU (Micro Processing Unit) or an ASIC (application-specific integrated circuit). The CPU 201 may also be an FPGA (Field Programmable Gate Array), etc.

[0050] ROM202 stores control programs such as the startup program for the information processing device 200. RAM203 is the main memory of the CPU201 and is used as work memory to temporarily store control programs or data when the CPU201 performs various control operations.

[0051] The storage device 204 is memory for storing applications, software, and various data. Examples of storage devices 204 include HDDs (Hard Disk Drives), SSDs (Solid State Drives), flash memory, and storage devices. For example, in the case of an LLM server 103, the storage device 204 stores the engine for large-scale language models.

[0052] The input / output interface 205 is an interface for transmitting output information to an output device such as the display unit 207. The display unit 807 can be a display or the like. Furthermore, the display unit 207 may be located outside the information processing device 200, as shown in Figure 2, or it may be located within the information processing device 200.

[0053] The input / output interface 205 is an interface for receiving input information from input devices such as the input unit 208. Examples of input devices 208 include keyboards, buttons, touch panels, operation keys, and pointing devices. Furthermore, the input unit 208 may be located outside the information processing device 200, as shown in Figure 2, or it may be located within the information processing device 200. The input / output interface 205 may also be configured as an input interface and an output interface.

[0054] Network I / F 206 is an interface for receiving data from other devices via network 105 and transmitting data generated or stored by the information processing device 200 to other devices via network 105.

[0055] The information processing device 200 may also be equipped with a media interface (not shown). The media interface reads programs and data stored in a storage device (not shown) connected by wire or wireless, and provides them to the CPU 201 via RAM 203. The CPU 201 can read the provided program from the storage device onto RAM 203 via the media interface and execute the read program. The storage device is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase Change Rewritable Disk), or a magneto-optical recording medium such as an MO (Magneto-Optical disk). Alternatively, the storage device may be a tape medium, a magnetic recording medium, or a semiconductor memory.

[0056] The CPU 201 of the information processing device 200 reads and executes these programs from the storage device, but as an alternative, these programs may be obtained from other devices via a predetermined communication network. Furthermore, it goes without saying that the information processing device 200 is provided with the necessary configurations to realize the information processing in this embodiment.

[0057] (An example of a user terminal configuration) Figure 3 shows an example of the configuration of a user terminal 101. The user terminal 101 has a communication unit 310, a control unit 320, and a storage unit 330. The control unit 320 includes a display control unit 321 and a processing unit 322. The storage unit 330 includes a web browser 331 and a notification application 332.

[0058] The web browser 331 sends a Hypertext Transfer Protocol (HTTP) request to the server indicated by the URL entered in the input area of ​​its window, and receives content data as a response. The web browser 331 then parses the received content data and can display the web page as a result of the parsing.

[0059] The notification application 332 can provide users with functions such as chatting and making online calls with other registered users via the notification server 104, as well as viewing notifications.

[0060] The communication unit 310 communicates with the platform server 102, notification server 104, and other devices via the network 105. The control unit 320 controls the entire user terminal 101.

[0061] The display control unit 321 controls the display of various information and various screens. For example, the display control unit 321 can display the window of the web browser 331 and the screen of the notification application 332. The processing unit 322 executes processing to realize the various functions of the user terminal 101.

[0062] (Example of a platform server configuration) Figure 4 shows an example of the configuration of the platform server 102. The platform server 102 has a communication unit 410, a control unit 420, and a storage unit 430. The control unit 420 includes an error detection unit 421, an acquisition unit 422, and a processing unit 423. The storage unit 430 includes a user database (DB) 431, an error log 432, an expense claim procedure application 433, and a prompt creation application 434.

[0063] User DB 431 contains information about users and groups such as companies that have contracts with the operator of platform server 102. Error log 432 is a log of errors that occur when system abnormalities in platform server 102 or errors caused by user operations occur. Error log 432 may be stored in user DB 431, linked to each user or group.

[0064] The group of applications running on the platform server 102 includes an expense claim procedure application 433 and a prompt generation application 434. The expense claim procedure application 433 is an application that provides services such as expense claim procedures. The expense claim procedure application 433 is used, for example, when expense claim procedures are carried out within a company.

[0065] The user can access the expense claim procedure application 433, which runs on the platform server 102, via the web browser 331 of the user terminal 101. In this case, the display unit 207 of the user terminal 101 will show the screen of the expense claim procedure application 433 as shown in Figure 5.

[0066] Figure 5 shows an example of the UI screen (hereinafter referred to as the expense application screen) of the expense application procedure application 433 provided by the platform server 102, which is displayed on the display unit 207 of the user terminal 101. The expense application screen 500 shown in Figure 5 includes a field 510 for basic input items. The expense application screen 500 may also include fields for other input items necessary for expense applications.

[0067] For example, when submitting an expense claim, the user submitting the claim performs input and selection operations on the expense claim screen 500. The user then confirms the input and selection items and submits the claim on the expense claim screen 500.

[0068] The prompt creation application 434 is an application that provides a service for creating prompts to send to the LLM server 103. Users can access the prompt creation application 434, which runs on the platform server 102, via the web browser 331 of the user terminal 101. In this case, the display unit 207 of the user terminal 101 will display the screen of the prompt creation application 434 as shown in Figure 6.

[0069] Figure 6 shows an example of the UI screen (hereinafter referred to as the prompt creation screen) of the prompt creation application 434 provided by the platform server 102, which is displayed on the display unit 207 of the user terminal 101. The prompt creation screen 600 shown in Figure 6 has a prompt input area 610, an error log input area 620, and an analysis result display area 630.

[0070] The prompt input area 610 displays prompts created by the user. The user can enter any string representing an instruction or setting as a prompt in the prompt input area 610. When the user operates the save button 611 provided on the prompt creation screen 600, the string entered in the prompt input area 610 is stored in the storage unit 430 of the platform server 102. Prompts may also be stored in the user DB 431, associated with each user or group. When an error occurs, the prompt created by the prompt creation application 434 is sent to the LLM server 103 along with its error log.

[0071] Error logs are displayed in the error log input area 620. Users can obtain any error log from a screen (not shown) provided by the platform server 102 and input that error log into the error log input area 620. When a user operates the send button 621 on the prompt creation screen 600, the prompt entered in the prompt input area 610 and the error log entered in the error log input area 620 are sent to the LLM server 103. The results of the analysis of the error log by the LLM server 103 according to the prompt are then displayed in the analysis result display area 630.

[0072] In this way, by using the prompt creation screen 600, users can create prompts and simulate the analysis results returned from the LLM server 103 using the created prompts and arbitrary error logs. In other words, by using the prompt creation screen 600, users can create and set prompts to obtain the desired analysis results.

[0073] Although the explanation described the expense application procedure application 433 and the prompt creation application 434 as being used via the web browser 331 of the user terminal 101, the method of use is not limited to this. For example, these applications may be installed on the user terminal 101 and used by the user.

[0074] The communication unit 410 communicates with the user terminal 101, the notification server 104, and other devices via the network 105. The control unit 420 controls the entire platform server 102. The error detection unit 421 detects errors (hereinafter referred to as "errors, etc.") caused by system abnormalities of the platform server 102 or user operations. The error detection unit 421 also generates a log of the detected errors, etc. and stores it in the storage unit 430 as an error log 432. When an error, etc. occurs, the acquisition unit 422 acquires the prompt stored in the storage unit 430 and its error log 432 and transmits it to the LLM server 103 via the communication unit 410.

[0075] The processing unit 423 performs processing to provide the user with the functions of each application running on the platform server 102. For example, when the processing unit 423 runs the expense application procedure application 433, it performs processing to display the expense application screen 500 on the user terminal 101 and related processing. Also, when the processing unit 423 runs the prompt creation application 434, it performs processing to display the prompt creation screen 600 on the user terminal 101 and related processing.

[0076] (Example of a notification server configuration) Figure 7 shows an example of the configuration of the notification server 104. The notification server 104 has a communication unit 710, a control unit 720, and a storage unit 730. The control unit 720 includes a processing unit 721. The storage unit 730 includes a user database 731 and a notification application 732.

[0077] User DB731 stores information about registered users. This user information includes the user ID and channel ID assigned to each user, as well as the user's name and other user-related information.

[0078] Notification application 732 is an application that provides chat services, network calling services, notification services, and other services. For example, notification application 732 can provide these services via notification application 332 installed on user terminal 101. Alternatively, notification application 732 can also provide these services via the web browser 331 on user terminal 101.

[0079] When a user uses these services, the display unit 207 of the user terminal 101 displays the screen of the notification application 732 as shown in Figure 8. Figure 8 is a diagram showing an example of the UI screen (hereinafter referred to as the notification screen) of the notification application 732 provided by the notification server 104, which is displayed on the display unit 207 of the user terminal 101. The notification screen 800 shown in Figure 8 has a message input area 810 and a notification display area 820.

[0080] Users can enter any string of characters as a message into the message input area 810. Users can notify any channel they belong to by entering a message into the message input area 810 and performing a notification operation. Other users belonging to that channel can then view the message that has been notified to that channel.

[0081] The message input area 810 is provided with a mention button 811 for using the mention function. The mention function is a function that indicates that a message has been notified to users or all users specified by this function among the users belonging to the channel. Users can notify the specified recipients (for example, some users or all users) by specifying the recipients with the mention button 811 and performing the notification operation. When a message is notified using the mention function, a notification item (not shown) indicating that a message has arrived is displayed on the recipient's notification screen 800 along with the message. Users can quickly notice that a message has arrived by looking at the notification item. Displaying a string indicating a specified user within the message or displaying a notification item along with the message notification on the notification screen 800 is sometimes referred to as mentioning.

[0082] The notification display area 820 shows the results of the error log analysis received by the notification server 104 from the platform server 102 via the API. In this way, if an error occurs on the platform server 102, the user can understand the details of the error from the notification screen 800. In addition to the results of the error log analysis, the notification display area 820 also displays messages (not shown) notified by other users.

[0083] The communication unit 710 communicates with the user terminal 101, the platform server 102, and other devices via the network 105. The control unit 720 controls the entire notification server 104. The processing unit 721 performs processing to provide the user with chat functions, network call functions, notification functions, and other functions, as well as processing related to these. The notification server 104 may also be configured to include applications corresponding to each service and processing units corresponding to those applications, and to execute the application corresponding to each service.

[0084] Regarding the LLM server 103, any configuration is acceptable as long as it can receive requests via API, analyze error logs according to prompts using a large-scale language model, and send the analysis results as a response. Therefore, a detailed explanation of the configuration of the LLM server 103 is omitted.

[0085] (Example of a process for notifying analysis results) Figure 9 is a sequence diagram showing an example of the process from when an error occurs on the platform server 102 until the results of the error log analysis are notified. Referring to Figure 9, the process from when the results of the error log analysis performed by the large-scale language model on the LLM server 103 are displayed on the display unit 207 of the user terminal 101 will be explained. In the example shown in Figure 9, an error occurs after the user has performed an input operation on the input fields of the expense application screen 500, and the error is caused by that input operation. The user performs an input operation on the expense application screen 500 displayed on the user terminal 101.

[0086] In S901, when the user terminal 101 detects a user operation to input and submit an input value on the expense application screen 500, it sends that input value to the platform server 102.

[0087] In S902, the platform server 102 detects errors in the input values ​​received in S901 that have been entered into the expense application screen 500. For example, an error occurs if a value that should not be entered into the expense application screen 500 is entered, or if a value that should be entered is not entered. Alternatively, an error occurs if an operation that should be performed by a user with the required privileges is performed by a user without the required privileges on the expense application screen 500.

[0088] In S903, the platform server 102 retrieves an error log 432 corresponding to the error detected in S902 from the storage unit 430. In S904, the platform server 102 retrieves a prompt previously set by the user from the storage unit 430.

[0089] In S905, the platform server 102 sends (requests) the error log 432 obtained in S903 and the prompt obtained in S904 to the LLM server 103. In S906, the LLM server 103 inputs the prompt and error log 432 received in S905 to the large-scale language model via the API. Then, in S906, the large-scale language model analyzes the error log 432 according to the prompt and outputs the analysis result. In S907, the LLM server 103 sends (responses) the analysis result of the error log obtained in S906 to the requesting platform server 102.

[0090] In S908, the platform server 102 sends (requests) the results of the error log analysis received in S907 to the notification server 104. At this time, the platform server 102 includes a pre-prepared channel ID in the request to the notification server 104.

[0091] In S909, the notification server 104 identifies the channel to which the error log analysis results should be notified from the channel ID included in the request received by S908 via the API, and sends a notification to the identified channel. The explanation of the response from the notification server 104 to the platform server 102 is omitted.

[0092] In S910, the user terminal 101 displays the error log analysis results on the notification screen 800 displayed on the display unit 207. By viewing the error log analysis results displayed on the notification screen 800, the user can understand the nature of the error that occurred.

[0093] Figure 10 is a flowchart showing an example of the processing performed by the CPU of the platform server 102 in this embodiment, as shown in the sequence diagram in Figure 9. The processing shown in the flowchart in Figure 10 is achieved by the CPU of the platform server 102 reading a program stored in ROM or a program stored in a storage device into RAM, and then executing the read program. The processing shown in the flowchart in Figure 10 starts after the platform server 102 is powered on and is repeatedly executed whenever an error occurs in the platform server 102.

[0094] In S1001, the CPU determines, using the error detection unit 421, whether or not it has detected an error or other issue caused by a system abnormality in the platform server 102 or by user operation. If the CPU determines in S1001 that it has detected an error or other issue (Yes), it proceeds to S1002. If the CPU determines in S1001 that it has not detected an error or other issue (No), it waits until it detects an error or other issue.

[0095] In S1002, the CPU uses the acquisition unit 422 to acquire an error log 432 corresponding to the error detected in S1001 from the storage unit 430, and proceeds to S1003. In S1003, the CPU uses the acquisition unit 422 to acquire a prompt previously set by the user from the storage unit 430, and proceeds to S1004.

[0096] In S1004, the CPU, via the communication unit 410, sends the error log 432 obtained in S1002 and the prompt obtained in S1003 to the LLM server 103, and proceeds to S1005.

[0097] In S1005, the CPU receives the error log analysis results from the LLM server 103 via the communication unit 410 and proceeds to S1006. In S1006, the CPU sends the error log analysis results received in S1005 to the notification server 104 via the communication unit 410, and the flow shown in Figure 10 ends.

[0098] Figure 11 is a flowchart showing an example of the processing performed by the CPU of the notification server 104 in this embodiment, as shown in the sequence diagram in Figure 9. The processing shown in the flowchart in Figure 11 is achieved by the CPU of the notification server 104 reading a program stored in ROM or a program stored in a storage device into RAM, and then executing the read program. The processing shown in the flowchart in Figure 11 starts after the notification server 104 is started up and is executed repeatedly each time a request is received from the platform server 102.

[0099] In S1101, the CPU receives the error log analysis results from the platform server 102 via the communication unit 710 and proceeds to S1102.

[0100] In S1102, the CPU, using the processing unit 721, identifies the channel based on the channel ID included in the request from the platform server 102 in S1101, and proceeds to S1103.

[0101] In S1103, the CPU, via the processing unit 721, notifies the channel identified in S1102 of the results of the error log analysis received in S1101.

[0102] According to this embodiment, when an error occurs, the user can be notified of the analysis results of the error log, which has been analyzed according to a pre-set prompt. This allows the user to understand the nature of the error that occurred. Furthermore, since the severity of the error is indicated in the analysis results of the error log, the user can quickly determine whether or not they should take immediate action on the error.

[0103] <<Second Embodiment>> In the first embodiment, an example was described in which the results of the error log analysis are notified when a system abnormality occurs in the platform server 102 or when an error occurs due to user operation. However, the severity of errors varies depending on the type of error. For example, some errors are of low severity, such as those caused by simple user mistakes, while others are of high severity and affect the entire system. For users, it may be preferable not to be notified of low-priority errors, but to be notified of more important errors. Therefore, in this embodiment, an example will be described in which, when a low-priority error occurs, the content of the error is not notified to the user as a result of the error log analysis.

[0104] The basic configurations of the user terminal 101, platform server 102, LLM server 103, and notification server 104 in this embodiment are the same as those of each device in the first embodiment, so a detailed explanation is omitted. The differences from the first embodiment will be described below.

[0105] Figure 12 shows an example of the configuration of the platform server 102 in this embodiment. The control unit 420 has an analysis unit 1200 in addition to the error detection unit 421, acquisition unit 422, and processing unit 423. The analysis unit 1200 determines whether or not to send the analysis results of the error log to the notification server 104. In this embodiment, the following prompts are used to make this determination. Some examples of prompts in this embodiment are shown below.

[0106] ------------------------------------------------------------------------------- ### Evaluation Rules (Applied in order of priority) 1. **Whitelist Check** -Error message containing "role required" →: Not checked -Error message containing "Registered role required" →: Not checked -Error message containing "If SelectableOption is FALSE, the workflow code must be "void"" →: Not checked -Error message containing "Items that violate the job authority regulations cannot be set" →: Not checked -Error message containing "No XXXXUser found to check role" →: Not checked -Error message containing "Audit_Events: String_of_HogeFugaVer did not find attribute for record Workflow_Approval" →: Not checked 2.**Check specific rules** -Module is "SQL" → :information_source: information -Module is "ServiceCenter" → :warning:Caution (Possible management system failure) -Module is "Timer" → Evaluated by delay time - "took more than N seconds," etc. → :warning:Caution -Failure and not re-executed → :rotating_light: Emergency -Module is "Integration" → External system integration error (depending on the content: :warning:Caution or :rotating_light:Emergency) 3. **Special Rules** -If "Invalid username or password" occurs: - Occurrences 5 times or less →:information_source: Information -6 or more occurrences (repeated occurrences) → :warning:Caution (Possible fraudulent login attempt or configuration error) 4. **Evaluation based on general standards (other than those mentioned above)** -:rotating_light:Critical: Persistent error or business disruption level (e.g., "Object reference...", "Violation of PRIMARY KEY...", "Process ... has been suspended") -:warning:Caution: Repeated performance degradation or partial malfunction (e.g., "Timeout" repeatedly occurring on the same screen) -:information_source:Info:Temporary errors that can be automatically recovered (e.g., one-time "Timeout expired", one-time "Invalid username or password") 5. **Promotion Rules Based on Frequency and Impact** - Single occurrence → Information (Info) - Repeated occurrences with the same function → Warning - Overall impact or permanent occurrence → Critical ------------------------------------------------------------------------------ Some of the above prompts correspond to the section describing the "evaluation rules" of the prompts described in the first embodiment. However, some of the above prompts differ from the section describing the "evaluation rules" in the first embodiment in the description of "whitelist verification". Specifically, some of the above prompts differ in that they describe "Error message contains "○○" →: Not checked". As a result, when a low-priority error included in the whitelist occurs, the [Urgency] item in the analysis result of the error log from the LLM server 103 will describe "Not checked" instead of ":information_source: Information (Info)". Below is an example of the output result that is output as an analysis result of the error log when the prompts and error log in this embodiment are input to the large-scale language model of the LLM server 103.

[0107] ----------------------------------------------------------------------------- [Urgency]: Not checked [Tenant Information]: XXXX [Time of occurrence / Error ID]: 2025-09-22 04:04:06 / 12345678-90ab-cdef-1234-567890abcdef [Error summary]: The error message "Registered role required" occurred, indicating that a registered role is required for the permission check process on a specific delegate approver settings screen. [Estimated scope of impact]: This issue occurs in the ScreenDataSetGetTargetEntities method within the MainFlow of the AssignedOverseerProfile module, affecting the processing of the permission delegation settings screen. [Recommended initial action]: This error is an informational error related to whitelisting. Check your permission settings and consider registering any missing roles. Check the logs in Service Center for more details. ----------------------------------------------------------------------------- The above error log analysis results differ from the error log analysis results described in the first embodiment in that the string "Not subject to check" is written in the [Urgency] field. The analysis unit 1200 analyzes the error log analysis results and can determine that if the string "Not subject to check" is included in the [Urgency] field, it will not send the error log analysis results to the notification server 104. The analysis unit 1200 can also determine that if the string "Information", "Caution", or "Urgent" is included in the [Urgency] field, it will send the error log analysis results to the notification server 104.

[0108] Thus, the whitelist functions as a filter to prevent notifications about low-priority errors. Hereafter, the instructions written to prevent notifications about low-priority errors in the prompt will be referred to as classification conditions. In other words, even if an error corresponding to an error message specified in the classification conditions occurs, the results of the error log analysis are not sent from the platform server 102 to the notification server 104, and the user is not notified.

[0109] The prompts in this embodiment are not limited to those described herein. For example, a description of classification conditions may be added to the prompt in addition to the description of "checking the whitelist." In this case, the types of error messages shown in the description of "checking the whitelist" and the types of error messages shown in the description of classification conditions may be the same or different.

[0110] Furthermore, the string used in the classification criteria is not limited to "Not Checked," but can be any specific string that is different from "Information," "Caution," and "Urgent," which indicate the severity of the error. In addition, the location where "Not Checked" is written in the classification criteria is not limited to the [Urgency] field; it can be written in any field of the output format.

[0111] Figure 13 is a flowchart showing an example of the processing performed by the CPU of the platform server 102 in this embodiment. The processing shown in the flowchart in Figure 13 starts after the platform server 102 is started and is repeatedly executed whenever an error occurs in the platform server 102. Note that the processing from S1001 to S1006 shown in Figure 13 is the same as the processing with the same reference numerals shown in Figure 10 in the first embodiment, so its explanation is omitted.

[0112] In S1301, the CPU uses the analysis unit 1200 to analyze the results of the error log received from the LLM server 103 in S1005, and proceeds to S1302. In S1302, the CPU uses the analysis unit 1200 to determine, based on the analysis results in S1301, whether or not to send the results of the error log received from the LLM server 103 in S1005 to the notification server 104.

[0113] If, in S1302, the CPU determines, via the analysis unit 1200, that it should send the analysis result of the error log received from the LLM server 103 in S1005 to the notification server 104 (Yes), then the process proceeds to S1006. If, in S1302, the CPU determines, via the analysis unit 1200, that it should not send the analysis result of the error log received from the LLM server 103 in S1005 to the notification server 104 (No), then the flow shown in Figure 12 is terminated. Note that the analysis method in S1301 and the determination method in S1302 are as described above, so their explanation is omitted.

[0114] According to this embodiment, the platform server 102 analyzes the results of the error log received from the LLM server 103 and determines whether or not to send it to the notification server 104 based on its contents. If the analysis results of the error log indicate a low-priority error, i.e., an error that is not subject to notification, the platform server 102 controls itself not to notify the user. If the analysis results of the error log indicate that the error is not subject to notification, the platform server 102 controls itself to notify the user. As a result, the user does not receive notifications for low-priority errors, but does receive notifications for higher-priority errors, allowing the user to save time and effort to deal with higher-priority errors.

[0115] <<Third Embodiment>> In the first embodiment, an example was described in which the results of error log analysis are notified to a channel identified by a channel ID. When the results of error log analysis are notified based on a channel ID, all users belonging to that channel can see the results. However, even if the results of error log analysis are notified, some users may not notice the notification. Therefore, in this embodiment, an example is described in which the results of error log analysis are notified and a notification item is displayed using the mention function of the notification application 732.

[0116] The basic configurations of the user terminal 101, platform server 102, LLM server 103, and notification server 104 in this embodiment are the same as those of each device in the first embodiment, so a detailed explanation will be omitted. The differences from the first embodiment will be explained below.

[0117] Figure 14 shows an example of the configuration of the notification server 104 in this embodiment. In this embodiment, the control unit 720 has a determination unit 1400 in addition to the processing unit 721. The determination unit 1400 determines whether or not a specific string is included in the analysis result of the error log received from the platform server 102. In this embodiment, there is a first string that has the function of mentioning all users belonging to the channel corresponding to the channel ID. In addition, there is a second string that has the function of identifying a specific user among the users belonging to the channel corresponding to the channel ID and mentioning that specific user. That is, if the determination unit 1400 determines that the specific string is included in the analysis result of the error log, the processing unit 721 can notify the channel corresponding to the channel ID of the analysis result while mentioning the user according to the type of specific string. In order to perform this determination, the following prompts are used in this embodiment. Below are some examples of prompts in this embodiment.

[0118] ------------------------------------------------------------------------------- ### Output Format Please output your Markdown file in the following order and labels. {Display the first string} [Urgency]: {Show urgency} [Tenant Information]: {Display Tenant (If Tenant=Specific_Items, convert to XXXX) (If Tenant=Hogehoge, convert to XXXX)} [Time of occurrence / Error ID]: {If ID and Time of Log are available → "yyyy-mm-dd hh:mm:ss" / xxx} {If only an ID is available, use "xxx"} {If only Time is available → "yyyy-mm-dd hh:mm:ss"} [Error Summary]: {Summarize the core of the error in 1-2 sentences} [Estimated scope of impact]: {Estimate what is happening in which function. If possible, specify the screen name or function name.} -If the Message contains the Entity name / Action name, please specify it. -If only the module name is provided, it will be labeled as "Processing within the relevant module". - No additional information available → "Cannot be confirmed" is displayed. [Recommended initial actions]: {Specific actions are provided for each error type} -DB connection error → "Check the operational status of the connection pool / DB" - Primary Key violation → "Check the app's duplicate registration logic" -Timer delay → "Check the logs of the relevant Timer and consider whether the processing time can be reduced." - External API error → "Check Integration logs and external system operational status" -Other unknowns → "Check detailed logs in Service Center" -Tenant / ID / Time of Log should only be output if they are included in the input. ------------------------------------------------------------------------------- Some of the above prompts correspond to the "output format" section of the prompt described in the first embodiment. However, some of the above prompts differ from the "output format" section in the first embodiment in that "Show first string" is added at the beginning. As a result, the first string is written at the beginning of the error log analysis results. If you want to mention a specific user, add "Show second string" at the beginning of the "output format" of the prompt. Below is an example of the output results that are output as a result of analyzing the error log when the prompts and error log in this embodiment are input to the large-scale language model of the LLM server 103.

[0119] ----------------------------------------------------------------------------- First string [Urgency]::information_source:Information (Info) [Tenant Information]: XXXX [Time of occurrence / Error ID]: 2025-09-22 04:04:06 / 12345678-90ab-cdef-1234-567890abcdef [Error summary]: The error message "Registered role required" occurred, indicating that a registered role is required for the permission check process on a specific delegate approver settings screen. [Estimated scope of impact]: This issue occurs in the ScreenDataSetGetTargetEntities method within the MainFlow of the AssignedOverseerProfile module, affecting the processing of the permission delegation settings screen. [Recommended initial action]: This error is an informational error related to whitelisting. Check your permission settings and consider registering any missing roles. Check the logs in Service Center for more details. ----------------------------------------------------------------------------- The error log analysis results described above differ from those described in the first embodiment in that a first string is written at the beginning. The determination unit 1400 analyzes the error log analysis results and determines that the first string or the second string is included. If the processing unit 721 mentions the string according to the string and notifies the channel of the error log analysis results. The determination unit 1400 analyzes the error log analysis results and determines that the first string or the second string is not included. If the processing unit 721 does not mention the string and notifies the channel of the error log analysis results.

[0120] Note that the location where the first or second string is written is not limited to the beginning of the error log analysis results. For example, the location where the first or second string is written may differ depending on the specifications of the API request in notification server 104. Of course, the location where the first or second string is written in the error log analysis results from LLM server 103 will be indicated by a prompt.

[0121] Figure 15 is a flowchart showing an example of the processing performed by the CPU of the notification server 104 in this embodiment. The processing shown in the flowchart in Figure 15 starts after the notification server 104 is started and is repeatedly executed each time a request is received from the platform server 102. Note that the processing from S1101 to S1103 shown in Figure 15 is the same as the processing with the same reference numerals shown in Figure 11 in the first embodiment, so its explanation is omitted.

[0122] In S1501, the CPU uses the determination unit 1400 to determine whether a specific string is included in the analysis result of the error log received in S1101. If the CPU determines in S1501 that the specific string is included in the analysis result of the error log received in S1101 (Yes), the process proceeds to S1502. If the CPU determines in S1501 that the specific string is not included in the analysis result of the error log received in S1101 (No), the process proceeds to S1103. In this case, in S1103, the CPU uses the processing unit 721 to notify the channel identified in S1102 of the analysis result of the error log received in S1101 without mentioning it.

[0123] In S1502, the CPU uses the determination unit 1400 to determine whether the specific string in S1501 is the first string or the second string. If the CPU determines in S1502 that the specific string in S1501 is the first string, it proceeds to S1503. If the CPU determines in S1502 that the specific string in S1501 is the second string, it proceeds to S1504.

[0124] In S1503, the CPU, via the processing unit 721, mentions all users belonging to the channel identified in S1102, notifies that channel of the results of the error log analysis received in S1101, and terminates the flow shown in Figure 15.

[0125] In S1504, the CPU, via the processing unit 721, mentions the user indicated by the second string, notifies the channel identified in S1102 of the results of the error log analysis received in S1101, and terminates the flow shown in Figure 15.

[0126] According to this embodiment, by instructing the system to write a specific string that serves as a mention in the prompt, the specific string is written in the error log analysis results from the LLM server 103. The notification server 104 then mentions the system based on the specific string and notifies the channel of the error log analysis results. As a result, the user can immediately recognize that a notification has been made, for example, by seeing the notification item.

[0127] <<Other Embodiments>> Furthermore, the embodiments described above can be combined as appropriate, as long as the processing content is not inconsistent.

[0128] For example, by instructing the system to write a specific string in the prompt that performs the function of mentioning a user depending on the type of error message, the system can change the way it mentions the user depending on the type of error and notify the system of the error log analysis results. For example, if the error message in the prompt indicates an error related to a system abnormality, the system can be instructed to write a first string. Or, if the error message in the prompt indicates an error caused by an operation by a user without the necessary privileges, the system can be instructed to write a second string indicating a user with the necessary privileges. This allows the system to change the user mentioned depending on the type of error message, i.e., the severity of the error, and notify the system of the error log analysis results.

[0129] Furthermore, the prompt instructs the user to write a specific string that serves as a mention for error messages other than those specified in the classification criteria. This allows for mentioning errors while simultaneously notifying the user of the analysis results of the error log for errors other than low-priority errors.

[0130] Furthermore, the notification server 104 may change the expression of the notification depending on the type of "Information," "Caution," and "Urgent" indicated in the [Urgency] item of the error log analysis result. For example, if the [Urgency] item of the error log analysis result is "Information," the notification server 104 may notify the error log analysis result using a string or symbol expressed in an inconspicuous tone (dark tone). Also, if the [Urgency] item of the error log analysis result is "Urgent," the notification server 104 may notify the error log analysis result using a string or symbol expressed in a conspicuous tone (bright tone).

[0131] Alternatively, depending on the API request specifications in notification server 104, a user ID may be used instead of a channel ID. In this case, the error log analysis results are notified to the user as a direct message, rather than to the channel. Such control is performed, for example, by platform server 102.

[0132] Furthermore, by changing the channel ID, the channel to which the results of error log analysis are notified can be changed. For example, if the error message indicates an error related to a system anomaly, the results of the error log analysis can be notified to the channel of the technical team for that system. Also, if the error message indicates an error related to expense claim procedures, the results of the error log analysis can be notified to the channel of the accounting team. Such control is performed, for example, on the platform server 102. [Explanation of symbols]

[0133] 101 User terminal 102 Platform Server 103 LLM Server 104 Notification Server< / screendatasetgetsurrogatehogetus>

Claims

1. A server system that exchanges information between an application server that runs an application, an LLM server with a large-scale language model, and a notification server that provides notification services, A detection unit that detects errors that occur in the aforementioned application server, A first acquisition unit acquires an error log corresponding to the error detected by the detection unit and a prompt that includes rules for classifying errors that may occur in the application server into multiple severity levels. A second acquisition unit acquires analysis information including information indicating the results of analyzing the error log on the LLM server based on the prompt, and information indicating the importance of the error detected by the detection unit among the multiple importance levels. A notification unit controls the notification of the analysis information acquired by the second acquisition unit to the user via the notification server. A server system characterized by having the following features.

2. The notification unit controls whether or not to notify the user of the analysis information via the notification server, based on the content of the analysis information. The server system according to feature 1.

3. The prompt includes a list of errors that belong to the first severity level among the multiple severity levels, The aforementioned notification unit, If the detection unit detects an error included in the list and the analysis information acquired by the second acquisition unit indicates a first importance level, control is performed to notify the analysis information. If the detection unit detects an error not included in the list, and the analysis information acquired by the second acquisition unit indicates a second importance level which is higher than the first importance level among the multiple importance levels, the unit performs control to notify the analysis information. The server system according to feature 2.

4. The first importance level is the lowest importance level among the multiple importance levels. The server system according to feature 3.

5. Errors belonging to the first severity level include errors caused by user actions. Errors belonging to the second severity level include errors related to system abnormalities of the application server. The server system according to feature 3 or 4.

6. The prompt includes an instruction to output in a predetermined output format, The second acquisition unit acquires the analysis information output by the LLM server based on the instruction. The server system according to any one of claims 1 to 4.

7. The predetermined output format includes at least an item indicating the severity of the error that occurred in the application server and an item indicating the content of the error. The server system according to claim 6, characterized in that it is as described above.

8. The predetermined output format further includes items indicating the actions the user should take to address the error. The server system according to claim 7, characterized in that it is a server system according to claim 7.

9. The server system according to claim 1, characterized in that the notification unit performs the control based on a pre-prepared identifier for identifying the user.

10. The aforementioned identifier is an identifier for identifying the channel to which the user belongs, The server system according to claim 9, characterized in that the notification unit notifies the channel of the analysis information acquired by the second acquisition unit via the notification server based on the identifier.

11. The notification server has a function to mention users belonging to the channel, The prompt includes an instruction to include a specific string that performs the function of mentioning, The server system according to claim 10, characterized in that when the second acquisition unit acquires the analysis information including the specific string, the notification unit notifies the channel of the analysis information and mentions the user indicated by the specific string.

12. The aforementioned specific string includes a first string that has the function of mentioning all users belonging to the channel, and a second string that has the function of mentioning a specific user among the users belonging to the channel. The aforementioned notification unit, When the second acquisition unit acquires the analysis information including the first string, it notifies the channel of the analysis information and mentions all users. The server system according to claim 11, characterized in that when the second acquisition unit acquires the analysis information including the second string, it notifies the channel of the analysis information and mentions the specific user.

13. The aforementioned mention displays a notification item indicating that the analysis information has been notified. The server system according to feature 11 or 12.

14. It has a storage unit that stores the prompt created by the user, The first acquisition unit acquires the prompt stored in the storage unit. The server system according to any one of claims 1 to 3.

15. The second acquisition unit acquires the analysis information based on the prompt created by the user as a simulation of the analysis information acquired from the LLM server and the error log entered by the user. The server system according to any one of claims 1 to 3.

16. The aforementioned application server is a platform server that runs multiple applications. The server system according to any one of claims 1 to 3.

17. An information processing device that is communicatively connected to an LLM server having a large-scale language model and a notification server that provides notification services, A detection unit for detecting errors that occur in the aforementioned information processing device, A first acquisition unit acquires an error log corresponding to the error detected by the detection unit and a prompt that includes a rule for classifying errors that may occur in the information processing device into multiple levels of severity. A second acquisition unit acquires analysis information including information indicating the results of analyzing the error log on the LLM server based on the prompt, and information indicating the importance of the error detected by the detection unit among the multiple importance levels. A notification unit controls the notification of the analysis information acquired by the second acquisition unit to the user via the notification server. An information processing device characterized by comprising:

18. A method for controlling an information processing device that is communicatively connected to an LLM server having a large-scale language model and a notification server that provides notification services, A detection step for detecting an error that occurred in the information processing device, A first acquisition step involves acquiring an error log corresponding to the error detected in the detection step, and a prompt that includes a rule for classifying errors that may occur in the information processing device into multiple levels of severity. A second acquisition step involves acquiring analysis information that includes information indicating the results of analyzing the error log on the LLM server based on the prompt, and information indicating the importance of the error detected in the detection step among the multiple importance levels. A notification step which controls the notification of the analysis information acquired in the second acquisition step to the user via the notification server. A control method characterized by having the following features.

19. An information processing device that is communicatively connected to an LLM server with a large-scale language model and a notification server that provides notification services, A detection unit for detecting errors that occur in the aforementioned information processing device, A first acquisition unit acquires an error log corresponding to the error detected by the detection unit and a prompt that includes a rule for classifying errors that may occur in the information processing device into multiple levels of severity. A second acquisition unit acquires analysis information including information indicating the results of analyzing the error log on the LLM server based on the prompt, and information indicating the importance of the error detected by the detection unit among the multiple importance levels. A notification unit controls the notification of the analysis information acquired by the second acquisition unit to the user via the notification server. A program characterized by being designed to function as such.

Images