A program processing method, device, computer device, and storage medium

Abstract

An embodiment of the present application discloses a program processing method, apparatus, computer device, and storage medium. The program processing method includes: obtaining a target version program, deploying the target version program to a first server sub-cluster in a server cluster, where the first server sub-cluster is a subset of the server cluster; obtaining first service data reported by the first server sub-cluster, where the first service data includes operation behavior chain data of a user on a page corresponding to the target version program; determining a first page health status of the target version program according to the first service data; determining a program deployment strategy for the server cluster according to the first page health status, and performing program deployment on the server cluster according to the program deployment strategy. By using the present application, the efficiency of version release can be improved, and the accuracy of the version release strategy can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular, to a program processing method, apparatus, computer device, and storage medium. Background Art

[0002] Today, with the rapid development of information technology, the Internet has become an important channel for people to quickly obtain, publish, and disseminate information, and it plays an important role in various aspects such as politics, economy, and life. Disseminating information on the Internet is mainly achieved through web pages, and users can obtain and disseminate information by viewing web pages. In order to enrich the functions and display effects of web pages, it is necessary to continuously upgrade and transform web pages.

[0003] In the existing version upgrade and transformation solution, developers deploy the new version program to all servers. The developers, as ordinary users, access the page corresponding to the new version program on their own terminal devices and perform various operations on the page, etc. If the developers do not encounter any problems during the trial, they consider the new version program to be stable; if they encounter page problems (such as page unresponsiveness, white screen, etc.), they consider the new version program to be unstable and roll back from the new version program to the historical version program.

[0004] The observation, evaluation, and decision-making carried out manually not only result in low efficiency of new version release, but also cause inaccurate version release strategies due to large subjective influence. Summary of the Invention

[0005] Embodiments of the present application provide a program processing method, apparatus, computer device, and storage medium, which can improve the efficiency of version release and the accuracy rate of version release strategies.

[0006] On the one hand, an embodiment of the present application provides a program processing method, including:

[0007] Obtain a target version program, and deploy the target version program to a first server sub-cluster in a server cluster, where the first server sub-cluster is a subset of the server cluster;

[0008] Obtain first service data reported by the first server sub-cluster, where the first service data includes operation behavior chain data of users on a page corresponding to the target version program;

[0009] Determine a first page health status of the target version program according to the first service data;

[0010] Determine a program deployment strategy for the server cluster according to the first page health status, and deploy a program to the server cluster according to the program deployment strategy.

[0011] One aspect of the embodiments of the present application provides a program processing device, including:

[0012] A first acquisition module, configured to acquire a target version program;

[0013] A first deployment module, configured to deploy the target version program to a first server sub-cluster in a server cluster, where the first server sub-cluster is a subset of the server cluster;

[0014] A second acquisition module, configured to acquire first service data reported by the first server sub-cluster, where the first service data includes operation behavior chain data of a user on a page corresponding to the target version program;

[0015] A status determination module, configured to determine a first page health status of the target version program according to the first service data;

[0016] A policy determination module, configured to determine a program deployment policy of the server cluster according to the first page health status;

[0017] A second deployment module, configured to perform program deployment on the server cluster according to the program deployment policy.

[0018] Wherein, the operation behavior chain data includes multiple user operation behaviors, and the first service data further includes a page access volume of each page where a user operation behavior is located;

[0019] The status determination module includes:

[0020] An overlay unit, configured to determine the number of behaviors of each user operation behavior according to the operation behavior chain data, obtain a first weight of each user operation behavior among the multiple user operation behaviors, and determine the behavior type of each user operation behavior;

[0021] A first determination unit, configured to determine the behavior health degree of each user operation behavior according to the first weight of each user operation behavior, the behavior type of each user operation behavior, the number of behaviors of each user operation behavior, and the page access volume;

[0022] The overlay unit is further configured to overlay the behavior health degrees of all user operation behaviors into a page health degree;

[0023] A second determination unit, configured to determine a first page health status of the target version program according to the page health degree.

[0024] Wherein, the first page health status includes a page normal status or a page abnormal status;

[0025] The second determination unit is specifically configured to:

[0026] If the page health degree is greater than the health degree threshold value, determine that the first page health state of the target version program is the page normal state;

[0027] If the page health degree is not greater than the health degree threshold value, determine that the first page health state of the target version program is the page abnormal state.

[0028] Among them, the first determination unit is specifically used for:

[0029] Determine the operation rate of the user operation behavior according to the number of user operation behaviors and the page access volume of the page where the user operation behavior is located;

[0030] Determine the user operation parent behavior that is forward adjacent to the user operation behavior on the operation behavior chain data, and determine the conversion rate of the user operation behavior according to the number of behaviors of the user operation behavior and the number of behaviors of the user operation parent behavior;

[0031] Determine the second weight of the user operation behavior according to the number of behaviors of the user operation behavior and the number of all user operation behaviors on the page where the user operation behavior is located;

[0032] Obtain the historical operation rate and historical conversion rate of the user operation behavior; the historical operation rate and historical conversion rate are determined according to the historical service data reported by the server cluster where the historical version program has been deployed;

[0033] Determine the behavior health degree of the user operation behavior according to the first weight, behavior type, operation rate, conversion rate, second weight, historical operation rate and historical conversion rate of the user operation behavior.

[0034] Among them, the number of conversion rates of the user operation behavior is N, and N is a positive integer;

[0035] The device further includes:

[0036] An abnormality determination module, configured to obtain N historical conversion rates, determine an abnormal conversion rate from the N conversion rates according to the N historical conversion rates and the N conversion rates, determine the abnormal operation behavior chain data corresponding to the abnormal conversion rate, determine an abnormal page according to the abnormal operation behavior chain data and the user operation behavior, and output the abnormal page.

[0037] Among them, the first page health state includes a page normal state or a page abnormal state;

[0038] The policy determination module is specifically used for:

[0039] If the health status of the first page is the normal page status, the continue deployment policy is determined as the program deployment policy of the server cluster;

[0040] If the health status of the first page is the abnormal page status, the rollback deployment policy is determined as the program deployment policy of the server cluster; the continue deployment policy and the rollback deployment policy are different from each other.

[0041] Among them, the program deployment policy includes the continue deployment policy;

[0042] The second deployment module includes:

[0043] The continue deployment unit is used to deploy the target version program to the second server sub-cluster in the server cluster according to the continue deployment policy; the second server sub-cluster is a subset of the server cluster, and the intersection between the first server sub-cluster and the second server sub-cluster is empty;

[0044] The acquisition unit is used to acquire the second service data reported by the first server sub-cluster and the second server sub-cluster, and determine the second page health status of the target version program according to the second service data; the second page health status includes the normal page status or the abnormal page status;

[0045] The acquisition unit is further used to, if the second page health status is the normal page status, deploy the target version program to the servers in the server cluster other than the first server sub-cluster and the second server sub-cluster.

[0046] Among them, it further includes:

[0047] The first rollback unit is used to, if the second page health status is the abnormal page status, roll back the first server sub-cluster and the second server sub-cluster from the target version program to the historical version program.

[0048] Among them, the program deployment policy includes the rollback deployment policy;

[0049] The second deployment module includes:

[0050] The second rollback unit is used to roll back the first server sub-cluster from the target version program to the historical version program according to the rollback deployment policy.

[0051] Among them, the second acquisition module is specifically used for:

[0052] Acquire the start timestamp when the target version program is deployed to the first server sub-cluster;

[0053] Determine the data reporting duration according to the starting timestamp;

[0054] If the data reporting duration reaches the duration threshold, obtain the service data associated with the target version program reported by the first server sub-cluster from the monitoring platform, and use the obtained service data as the first service data.

[0055] Among them, the first deployment module is specifically used for:

[0056] Submit the target version program to the container platform, so that the container platform selects the first server sub-cluster from the server cluster, and the container platform deploys the target version program to the first server sub-cluster.

[0057] One aspect of the embodiments of the present application provides a computer device, including a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the methods in the above embodiments.

[0058] One aspect of the embodiments of the present application provides a computer storage medium. The computer storage medium stores a computer program. The computer program includes program instructions. When the program instructions are executed by the processor, the methods in the above embodiments are executed.

[0059] One aspect of the embodiments of the present application provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions. The computer instructions are stored in a computer-readable storage medium. When the computer instructions are executed by the processor of the computer device, the methods in the above embodiments are executed.

[0060] The present application does not require manual participation. The terminal device automatically determines the deployment strategy according to the running situation of the new version program online, avoiding the interference of subjective factors brought by manual decision-making, and improving the efficiency of version release and the accuracy of the version release strategy; furthermore, only some servers are deployed with the new version program. Even if the new version program is unstable, the impact on the entire business service is limited, which can reduce the version iteration risk; further, the present application uses the operation behavior chain data of the user on the page as the basis for verifying the health status of the page. The operation behavior chain data can completely trace the user's operation process on the page. Therefore, based on the operation behavior chain data, the health status of the page can be determined more accurately, which can further improve the accuracy of the deployment strategy. Description of the Drawings

[0061] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0062] Figure 1 It is a system architecture diagram of a program processing provided by an embodiment of the present application;

[0063] Figure 2 It is a schematic diagram of a program processing scenario provided by an embodiment of the present application;

[0064] Figure 3 It is a flow schematic diagram of a program processing method provided by an embodiment of the present application;

[0065] Figures 4a - 4d It is a schematic diagram of operation behavior chain data provided by an embodiment of the present application;

[0066] Figure 5 It is an explanatory schematic diagram of a first weight provided by an embodiment of the present application;

[0067] Figures 6a - 6c It is an explanatory schematic diagram of a user behavior type provided by an embodiment of the present application;

[0068] Figure 7 It is a flow schematic diagram of a program processing method provided by an embodiment of the present application;

[0069] Figure 8 It is a structural schematic diagram of a program processing device provided by an embodiment of the present application;

[0070] Figure 9 It is a structural schematic diagram of a computer device provided by an embodiment of the present invention. Detailed implementation manners

[0071] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0072] Please refer to Figure 1, which is a system architecture diagram for program processing provided by an embodiment of this application. This application relates to a server cluster 10f and a terminal device cluster. The terminal device cluster may include multiple terminal devices, such as terminal device 10a, terminal device 10b, terminal device 10c, etc. The server cluster 10f may include server sub-cluster 10d and server sub-cluster 10e, and each server sub-cluster may contain multiple servers.

[0073] When a new version of a program needs to be released on the server, the new version of the program is deployed to server sub-cluster 10d. The servers in server sub-cluster 10d run the new version of the program and provide services for the terminal device cluster. During the process of users in the terminal device cluster trying out the new version of the program, server cluster 10d reports user behavior chain data, and determines the page health status of the new version of the program based on the reported user behavior chain data. If the page health status is the normal page status, the new version of the program is deployed to server sub-cluster 10e, so that all servers are deployed with the new version of the program, that is, this version iteration is successful; otherwise, if the page health status is the abnormal page status, the new version of the program is rolled back to the historical version of the program in server sub-cluster 10d, that is, this version iteration fails. Although this version iteration fails, the overall impact on the service is small, and the version iteration risk is minimized to the greatest extent. Subsequently, developers can make targeted corrections to the new version of the program based on the user behavior chain data and wait for the next version iteration.

[0074] Among them, Figure 1 The terminal devices 10a, 10b, 10c, etc. shown may be mobile phones, tablet computers, laptop computers, handheld computers, mobile internet devices (MIDs), wearable devices, or other intelligent devices that can perform page operations.

[0075] Please refer to Figure 2 , which is a schematic diagram of a program processing scenario provided by an embodiment of this application. As Figure 2As shown, there are currently 5 servers, and historical version programs are deployed on all 5 servers. The historical version programs can provide corresponding services for user terminals. For example, page access services, communication message sending servers, payment services, and so on. When a version upgrade is required, a new version program is obtained and submitted to the container platform, which is used to uniformly schedule and manage these 5 servers. The container platform selects the servers for deploying the new version program from these 5 servers based on a predetermined selection strategy. Suppose the container platform selects server cluster 20b for deploying the new version program. Among them, the selection strategy can be to select k servers with the least load for deploying the new version program. The container platform will deploy the new version program to server cluster 20b. At this time, the new version program is running in server cluster 20b, and the remaining 3 servers are running historical version programs.

[0076] Whether it is a server running a historical version program or a server running a new version program, during the service provision process, it will report the operation behavior chain data of users on the page to the monitoring platform. This operation behavior chain data represents the complete operation behavior process of users on the page, and each operation behavior chain data includes at least one user operation behavior. Obtain the operation behavior chain data reported by server cluster 20b, and analyze the behavior health of multiple user operation behaviors based on this operation behavior chain data.

[0077] Among them, the behavior health can be determined based on the ratio between the number of user operation behaviors and the page access volume of the page where the user operation behavior is located. The behavior health can be determined based on the ratio between the number of user operation behaviors and the number of user operation parent behaviors of the user operation behavior. If user operation behavior A and user operation behavior B belong to the same operation behavior chain data, and user operation behavior B occurs after user operation behavior A, it means that user operation behavior A is the user operation parent behavior of user operation behavior B. The behavior health can also be determined based on the ratio between the number of user operation behaviors and all user operation behaviors on the page where the user operation behavior is located.

[0078] In the above manner, calculate the behavior health of each user operation behavior involved in the operation behavior chain data respectively, and superimpose all behavior healths into the page health. If the page health is greater than the preset health threshold, it is considered that the page state of the page corresponding to the new version program is in a normal page state; otherwise, if the page health is not greater than the preset health threshold, it is considered that the page state of the page corresponding to the new version program is in an abnormal page state.

[0079] If it is determined that the page status is in the normal page state, it indicates that the performance of the new version program is stable, and the new version program can be continuously deployed on the remaining servers. That is, the container platform deploys the new version program to the remaining 3 servers. At this point, the new version program runs on all 5 servers, and the new version program running on these 5 servers can provide corresponding services. Then, the upgrade process of the server from the historical version program to the new version program is completed.

[0080] If it is determined that the page status is in the abnormal page state, it indicates that the performance of the new version program is unstable and there are program errors. Roll back from the new version program to the historical version program in server cluster 20b, that is, delete the new version program in server cluster 20b and continue to run the historical version program. At this time, the historical version program is still running on all 5 servers.

[0081] Subsequently, developers can determine abnormal user operation behaviors based on the behavior health of each user operation behavior, and then locate the abnormal page where the abnormal user operation behavior is located. They can make targeted modifications to the new version program based on the abnormal page.

[0082] This application uses the operation behavior chain data of the user on the page as the basis for verifying the page health status. The operation behavior chain data can completely trace the user's operation process on the page. Therefore, based on the operation behavior chain data, the page health status can be determined more accurately, which can improve the accuracy of the deployment strategy. Further, this application determines abnormal user operation behaviors based on the health of each user operation behavior, and can quickly and accurately locate the abnormalities on the page.

[0083] Among them, the specific process of obtaining the target version program (such as the new version program in the above embodiment), deploying the target version program to the first server sub-cluster (such as server cluster 20b in the above embodiment), obtaining the first business data (such as the operation behavior chain data reported by server cluster 20b in the above embodiment), and determining the first page health status (such as the page status in the above embodiment) can be referred to in the following Figures 3 - 7 corresponding embodiment.

[0084] Please refer to Figure 3 , which is a schematic flowchart of a program processing method provided by an embodiment of this application. The following embodiment is described with the program deployment device as the execution subject. The program deployment device is installed with a continuous deployment integration system, and the continuous deployment integration system includes an intelligent continuous deployment system. This intelligent continuous deployment system is integrated in the continuous deployment integration system as a plugin. The intelligent continuous deployment integration system is used to deploy software to the production environment in an automated, frequent, and continuous manner, enabling the software product to develop rapidly through iteration.

[0085] The program processing method may include the following steps:

[0086] Step S101 : obtaining a target version program, and deploying the target version program to a first server sub-cluster in a server cluster, where the first server sub-cluster is a subset of the server cluster.

[0087] Specifically, the program deployment device obtains the target version program to be deployed (as described above Figure 2 The target version of the program is submitted to the container platform (Kubernetes platform). The container platform is used to manage containerized applications on multiple servers. The goal of the container platform is to make the deployment of containerized applications simple and powerful. Kubernetes provides a mechanism for application deployment, planning, updating, and maintenance.

[0088] The container platform selects from the server cluster (as described above) according to the preset selection strategy Figure 2 The server sub-cluster (referred to as the first server sub-cluster, as described above) for deploying the target version program is selected from the five servers in the corresponding embodiment. Figure 2 The server cluster 20b) in the corresponding embodiment, wherein the preset selection strategy may be a minimum load strategy, that is, the container platform selects k servers with the minimum load from the server cluster to form the first server sub-cluster. The preset selection strategy may also be a sequential strategy, that is, the servers in the server cluster are used to deploy the new version of the program in sequence. For example, the servers with serial numbers 1, 2, and 3 are selected this time to deploy the target version of the program, and the servers with serial numbers 4, 5, and 6 are selected next time to deploy the next version of the target version of the program.

[0089] The number of servers in the first server sub-cluster may be 10% of the server cluster.

[0090] Among them, the servers in the server cluster can be deployed with historical versions of the program. The program deployed on the server in this application means that the program is installed and runs on the server, that is, at this time, the servers in the server cluster are running historical versions of the program.

[0091] The container platform deploys the target version program in the first server sub-cluster. After the deployment, the target version program runs in the first server sub-cluster, and the historical version program runs in the servers other than the first server sub-cluster in the server cluster.

[0092] Step S102: acquiring first business data reported by the first server sub-cluster, wherein the first business data includes operation behavior chain data of a user on a page corresponding to the target version program.

[0093] Specifically, whether it is the server running the historical version program or the server running the target version program, the server will report business data to the monitoring platform. The business data includes all the behavior data of users on the corresponding pages of the historical version program / target version program. Record the timestamp when the target version program is deployed to the first server sub-cluster (referred to as the start timestamp). According to the start timestamp, the data reporting duration is statistically calculated. If the data reporting duration reaches the preset duration threshold, the program deployment device can obtain the business data associated with the target version program reported by the first server sub-cluster from the monitoring platform, and use the business data obtained from the monitoring platform as the first business data.

[0094] Among them, the first business data includes the operation behavior chain data of users on the corresponding page of the target version program. The number of operation behavior chain data can be multiple. Each operation behavior chain data includes at least one user operation behavior and one operation result, and each operation behavior chain data ends with the operation result. The operation behavior chain data can represent the complete operation process of the user on the page and the sequence of multiple user operation behaviors.

[0095] Among them, the page in this application can specifically refer to a Web page, and this Web page can specifically be the personalized dressing Web page in an instant messaging application.

[0096] For example, in the page corresponding to the user's target version program, the user clicks the button "I want to buy", and a window pops up. The window includes a prompt message: "Are you sure you want to buy?" and buttons "Yes" and "No"; the user then clicks the button "Yes"; a password input window pops up, the user enters a password in the password input window, and clicks the button "OK", and a window pops up to prompt that the purchase is successful. In this operation behavior chain data, 3 user operation behaviors are involved. The first user operation behavior is that the user clicks the button "I want to buy", the second user operation behavior is that the user clicks the button "Yes", and the last user operation behavior is that the user enters a password and clicks the button "OK", and the operation result is: the purchase is successful.

[0097] Please refer to Figures 4a - 4d , which is a schematic diagram of an operation behavior chain data provided by an embodiment of this application. On the page shown in Figure 4a , the user clicking the button "Click to buy a membership" can be the user operation behavior ①. The program deployment device responds to the user operation behavior ① and jumps to the page shown in Figure 4b . The user clicking the button "Confirm" can be the user operation behavior ②. The program deployment device responds to the user operation behavior 2, and there may be 2 operation results. One of the operation results ③ corresponds to the page shown in Figure 4c , and this page indicates that the user has successfully purchased a membership; the other operation result ④ corresponds to the page shown in Figure 4dThe page shown, which indicates that the user's membership purchase failed. In summary, Figures 4a - 4c operation behavior chain data can be generated: user operation behavior ① → user operation behavior ② → operation result ③; Figures 4a - 4b and Figure 4d operation behavior chain data can be generated: user operation behavior ① → user operation behavior ② → operation result ④. Among them, user operation behavior ① is the parent behavior of user operation behavior ②, and user operation behavior ② is the child behavior of user operation behavior ①.

[0098] Step S103, determine the first page health status of the target version program according to the first service data.

[0099] Specifically, the first service data further includes the page access volume of all pages corresponding to the target version program. The program deployment device counts the number of behaviors of multiple user operation behaviors included in the operation behavior chain data according to the above operation behavior chain data. Each user operation behavior has a uniquely corresponding page. Obtain the page access volume of the page where each user operation behavior is located from the first service data.

[0100] The program deployment device obtains the weight of each user operation behavior among all user operation behaviors (referred to as the first weight). The first weight can represent the importance degree of the user operation behavior among all user operation behaviors. The sum of the first weights of all user operation behaviors is equal to 1. For example, on a message publishing page, the user operation behavior of clicking the "Send" button has a greater impact on the business than the user operation behavior of clicking an advertisement on the page. Therefore, the first weight of the user operation behavior of clicking the "Send" button is greater than the first weight of the user operation behavior of clicking an advertisement on the page.

[0101] Please refer to Figure 5 , Figure 5 which is a schematic diagram for explaining a first weight provided by an embodiment of the present application. Figure 5 The page shown is a message publishing page. On this page, the user can enter text content. After clicking the button "Confirm and Send", the user's friends can see this text content. In Figure 5 the page shown, there can be 2 user operation behaviors. One user operation behavior is clicking the button "Confirm and Send", and the other user operation behavior is clicking the recommended content on the page. The recommended content can be advertisement content. Since Figure 5 the page shown is a message publishing page, clicking the button "Confirm and Send" is a necessary operation step in page design interaction, and clicking the recommended content on the page is not a necessary operation step in page design interaction. Therefore, the first weight of the user operation behavior of clicking the button "Confirm and Send" must be greater than the first weight of the user operation behavior of clicking the recommended content on the page.

[0102] The program deployment device obtains the behavior type of each user operation behavior. The behavior type includes positive behavior types or negative behavior types. User operation behaviors where page anomalies are a necessary condition for the occurrence of user operation behaviors are called user operation behaviors belonging to the negative behavior type. User operation behaviors other than those belonging to the negative behavior type all belong to the positive behavior type.

[0103] Please refer to Figures 6a - 6c , which is a schematic diagram for explaining a type of user behavior provided by an embodiment of the present application. Figure 6a The page shown is a message publishing page. After the user clicks the button "Confirm and Send" on the page, if the message is successfully sent, it will jump to Figure 6b the page shown; conversely, if the message sending fails, it will jump to Figure 6c the page shown. Since the user clicks the button "Confirm and Send" on the page shown in Figure 6a and clicks the button "Confirm" on the page shown in Figure 6b are normal interaction behaviors of the page and are also considered expected behaviors; while clicking the button "Retry" on the page shown in Figure 6c is an interaction path that will be entered only because some anomaly has occurred on the page, that is, page anomaly is a necessary condition for the user operation behavior of clicking the button "Retry". Therefore, the behavior types to which the user operation behavior of clicking the button "Confirm and Send" and the user operation behavior of clicking the button "Confirm" belong are both positive behavior types; the behavior type to which the user operation behavior of clicking the button "Retry" belongs is a negative behavior type.

[0104] So far, the program deployment device has obtained the first weight of each user operation behavior, the behavior type of each user operation behavior, the behavior quantity of each user operation behavior, and the page access volume of multiple pages. The program deployment device determines the behavior health degree of each user operation behavior based on the first weight, behavior type, behavior quantity, and page access volume of each user operation behavior above. This behavior health degree is used to represent the health degree of the user operation behavior. The program deployment device adds up the behavior health degrees of all user operation behaviors to obtain the page health degree. Obtain a preset health degree threshold. If the page health degree is greater than the health degree threshold, it is determined that the first page health state of the target version program is the page normal state; if the page health degree is less than or equal to the health degree threshold, it is determined that the first page health state of the target version program is the page abnormal state.

[0105] Next, taking a user operation behavior as an example, it is explained how to determine the behavior health degree of the user operation behavior based on the first weight, behavior type, behavior quantity, and page access volume of the user operation behavior:

[0106] The ratio obtained by dividing the number of occurrences of a user operation behavior by the page view volume of the page where the user operation behavior is located is called the operation rate of the user operation behavior. The operation rate indicates the proportion of users who will perform a certain user operation behavior on a certain page. Generally speaking, if the interaction of the page remains unchanged, the operation rate of a user operation behavior can be considered stable; if the operation rate of a certain user operation behavior fluctuates violently, it is often caused by page failures. For example, an increase in the click-through rate of the retry button when sending a message fails may indicate an increase in the failure rate of the message sending service; a decrease in the click-through rate of the payment button in the payment pop-up window may indicate a problem with the loading of the payment pop-up window, etc.

[0107] Determine the user operation behavior that is forward adjacent to the user operation behavior (referred to as the user operation parent behavior) in the operation behavior chain data. The user operation behavior B that is forward adjacent to the user operation behavior A means that in the operation behavior chain data, the user operation behavior B is located before the user operation behavior A, and there are no other user operation behaviors between the user operation behavior B and the user operation behavior A; from the user's perspective, it means that after performing the user operation behavior B, the user operation behavior A is performed, and no other user operation behaviors are performed during the execution of these two user operation behaviors. The ratio obtained by dividing the number of occurrences of the user operation behavior by the number of occurrences of the user operation parent behavior is called the conversion rate of the user operation behavior. The conversion rate indicates how many users will continue to perform the user operation behavior B after performing the user operation behavior A in an operation behavior chain data. Generally speaking, if the interaction of the page remains unchanged, the conversion rate of a user operation behavior in an operation behavior chain data can be considered stable. If the conversion rate of a certain user operation behavior in an operation behavior chain data fluctuates violently, it is also often caused by page failures. For example, on a page for setting up a dress-up, when the user clicks on the dress-up details, a dress-up pop-up window will be called up, and clicking the button "Set" can complete the dress-up. If the conversion rate of the user operation behavior of clicking the button "Set" decreases, it may indicate a decrease in the success rate of loading the dress-up pop-up window.

[0108] It should be noted that a user operation behavior may exist in more than one operation behavior chain data. If a user operation behavior exists in multiple operation behavior chain data, then the number of user operation parent behaviors of this user operation behavior is multiple, and correspondingly, there are multiple conversion rates. Each conversion rate represents the conversion rate of the user operation behavior in an operation behavior chain data. Subsequently, there are also multiple corresponding behavior health degrees for this user operation behavior.

[0109] Since the conversion rate is determined by the number of actions of the user operation parent action of the user operation behavior, for any action chain data of the user operation, there will always be a first user operation behavior. The first user operation behavior has no user operation parent action. At this time, directly set the conversion rate of the first user operation behavior to 100%.

[0110] The ratio of the number of actions of the user operation behavior to the sum of the number of actions of all user operation behaviors on the page where the user operation behavior is located is called the second weight of the user operation behavior. The second weight represents the importance of a certain user operation behavior on the page. The higher the second weight, the greater the impact on the behavior health.

[0111] Obtain the historical operation rate and historical conversion rate of the user operation behavior. Among them, the historical operation rate and historical conversion rate are determined according to the historical business data reported by the server cluster deployed (running) with the historical version program. The method of determining the historical operation rate is the same as the method of determining the operation rate according to the first business data described above, and the method of determining the historical conversion rate is the same as the method of determining the conversion rate according to the first business data described above.

[0112] At this point, the program deployment device has obtained the operation rate, conversion rate, second weight, first weight, behavior type, historical operation rate and historical conversion rate of a user operation behavior.

[0113] Calculate the behavior health h of the user operation behavior i based on the following formula (1) i :

[0114]

[0115] Among them, α i is the operation rate of the user operation behavior i, β i is the conversion rate of the user operation behavior i, W i is the first weight of the user operation behavior i, P i is the second weight of the user operation behavior i, α 0i is the historical operation rate of the user operation behavior i, β 0i is the historical conversion rate of the user operation behavior i. If the behavior type of the user operation behavior i is a positive behavior type, then Q i = 1. If the behavior type of the user operation behavior i is a negative behavior type, then Q i = -1.

[0116] Analyzing formula (1), it can be known that the behavior health of the user operation behavior is to measure the degree of change of the user operation behavior from the historical version program to the target version program. If the change is drastic, the behavior health is small, indicating that the target version program is unstable; on the contrary, if the change is drastic, the behavior health is large, indicating that the target version program is stable.

[0117] By adding up the behavior health levels of all user operation behaviors, the page health level can be obtained. The calculation formula for the page health level is as shown in the following formula (2):

[0118] H = ∑ i h i (2)

[0119] It should be noted that due to version iteration, there may be some user operation behaviors that only exist in the target version program but not in the historical version program. The behavior health levels of these user operation behaviors can be directly set to a preset value (for example, directly set to 1).

[0120] Optionally, as described above, user operation behaviors may exist in more than one piece of user operation behavior chain data. If a user operation behavior exists in multiple pieces of user operation behavior chain data, then the number of user operation parent behaviors of this user operation behavior is multiple, and correspondingly, the conversion rates are also multiple. Each conversion rate represents the conversion rate of the user operation behavior on one piece of operation behavior chain data. Subsequently, there are also multiple behavior health levels for this user operation behavior. If the number of conversion rates of a certain user operation behavior is N, where N is a positive integer. Obtain N historical conversion rates corresponding to these N conversion rates respectively, and select abnormal conversion rates from the N conversion rates according to the N historical conversion rates, where the difference between the abnormal conversion rate and the historical conversion rate is greater than the difference threshold. Take the operation behavior chain data corresponding to the abnormal conversion rate as the abnormal operation behavior chain data, and determine the abnormal page based on the abnormal operation behavior chain data and this user operation behavior. The abnormal page can be the page where this user operation behavior is located, or the page where the user operation behavior adjacent to the user operation behavior forward on the abnormal operation behavior chain data is located, or the page where the user operation behavior adjacent to the user operation behavior backward on the abnormal operation behavior chain data is located.

[0121] Optionally, if the difference between the historical operation rate and the operation rate of a user operation behavior is greater than the difference threshold, then determine the page where this user operation behavior is located as the abnormal page and output the abnormal page.

[0122] The output abnormal page can help developers quickly and accurately locate the abnormal position in the target version program, and subsequently, targeted modifications can be made to the target version program.

[0123] Step S104, determine the program deployment strategy of the server cluster according to the first page health status, and deploy the program to the server cluster according to the program deployment strategy.

[0124] Specifically, as can be seen from the above, the health status of the first page can be a normal page status or an abnormal page status. If the health status of the first page is a normal page status, the program deployment device will continue to deploy the policy as the program deployment policy for the server cluster. Among them, the continue deployment policy means continuing to deploy the target version program to the server cluster.

[0125] If the health status of the first page is an abnormal page status, the program deployment device will roll back the deployment policy as the program deployment policy for the server cluster. Among them, the rollback deployment policy means rolling back the deployment of the target version program in the first server sub-cluster to the historical version program, that is, the first server sub-cluster no longer runs the target version program and instead starts running the historical version program. It can also be considered that the version upgrade from the historical version program to the target version program fails. The program deployment device can generate a notification message for the failed version iteration.

[0126] If the program deployment policy is the rollback deployment policy, the program deployment device rolls back the first server sub-cluster from the target version program to the historical version program according to the rollback deployment policy. After the rollback, the historical version program is running on each server in the server cluster. Rollback means restoring the program in the server to the previous version. In this application, it is to revoke the target version program deployed in the first server sub-cluster and restore the historical version program to run in the first server sub-cluster.

[0127] If the program deployment policy is the continue deployment policy, the program deployment device notifies the container platform to continue selecting a second server sub-cluster from the remaining servers according to the continue deployment policy. Of course, the intersection between the second server sub-cluster and the first server sub-cluster is empty (the number of servers in the second server sub-cluster can be 40% of the server cluster). The container cloud platform deploys the target version program to the second server sub-cluster. At this time, the target version program is running in both the first server sub-cluster and the second server sub-cluster. Whether it is the server running the historical version program or the server running the target version program, the server will report the business data to the monitoring platform. The business data includes all the behavior data of the user on the page corresponding to the historical version program / target version program. Record the timestamp when the target version program is deployed to the second server sub-cluster (referred to as the auxiliary timestamp). According to the auxiliary timestamp, calculate the auxiliary duration of data reporting. If the data reporting auxiliary reaches the preset duration threshold, the program deployment device can pull the business data associated with the target version program reported by the first server sub-cluster from the monitoring platform and pull the business data associated with the target version program reported by the second server sub-cluster from the monitoring platform, and use the business data obtained from the monitoring platform as the second business data.

[0128] Among them, the second service data includes the auxiliary operation behavior chain data of the user on the page corresponding to the target version level. The auxiliary operation behavior chain data includes the operation behavior chain data mentioned above. The number of auxiliary operation behavior chain data can be multiple. Each auxiliary operation behavior chain data includes at least one user operation behavior and one operation result. Each auxiliary operation behavior chain data ends with an operation result. The auxiliary operation behavior chain data can represent the complete operation process of the user on the page and the sequence of multiple user operation behaviors.

[0129] The program deployment device determines the second page health status of the target version program according to the second service data. The calculation process for determining the second page health status is the same as the calculation process for determining the first page health status according to the first service data mentioned above, except that the processed data is different.

[0130] Similarly, the second page health status can be the page normal status or the page abnormal status. If the second page health status is the page normal status, the program deployment device notifies the container platform to deploy the target version program to the servers in the server cluster other than the first server sub-cluster and the second server sub-cluster. At this time, each server in the server cluster has deployed the target version program, that is, the upgrade process from the historical version program to the target version program is completed.

[0131] If the second page health status is the page abnormal status, it means that the performance of the target version program is unstable. Then the program deployment device notifies the container platform to roll back the first server sub-cluster and the second server sub-cluster from the target version program to the historical version program. After the rollback, the historical version program is deployed on each server in the server cluster. It can also be considered that the version upgrade from the historical version program to the target version program fails. The program deployment device can generate a notification message for the version iteration failure.

[0132] This application proposes a brand-new calculation method for page health, which can be used to measure the stability of the new version program in version iteration and enrich the methods for measuring program stability. Moreover, the behavior health degree in this application represents the change degree of the user operation behavior from the historical version program to the target version program, accurately measuring the stability of the new version program and further ensuring the accuracy of the program deployment strategy.

[0133] Please refer to Figure 7 , which is a schematic flowchart of a program processing method provided by an embodiment of this application. The program processing includes the following steps:

[0134] Step S201, when the developer merges the target version program from the branch to the main trunk, it triggers the intelligent continuous deployment system to start the new version release.

[0135] Specifically, the program processing method involved in this application can be integrated into an intelligent continuous deployment system, which can be integrated into a Continuous Integration system as a plugin of the Continuous Integration system. The target version program can specifically be the program corresponding to a web page.

[0136] When a developer merges the target version program from a branch into the main trunk, the Continuous Integration system triggers the intelligent continuous deployment system to start the version release.

[0137] Step S202, the intelligent continuous deployment system obtains the target version program and starts the health calculation system in the intelligent continuous deployment system.

[0138] Step S203, the intelligent continuous deployment system deploys the target version program to 10% of the formal environment.

[0139] Specifically, the formal environment refers to the server cluster that bears the access of external network users. The process of updating the code of the formal environment to the new version code is called deployment (or release). The intelligent continuous deployment system submits the target version program to the container platform, and the container platform deploys the target version program to 10% of the server sub-cluster of the full-scale servers.

[0140] Step S204, 10% of the server sub-cluster reports business data to the monitoring platform.

[0141] Specifically, the monitoring platform bears the business data reported by the Web page and provides it for the health calculation module to query. It has the functions of storing data and providing API to query data. Extending, outside the release process, the monitoring platform also monitors the daily business data of the Web page and should have the ability of active alarm.

[0142] Step S205, the health calculation system in the intelligent continuous deployment system pulls the business data reported by 10% of the server sub-cluster from the monitoring platform.

[0143] Step S206, the health calculation system calculates the page health based on the business data. If the page health is greater than the preset threshold, step S208 is executed; if the page health is not greater than the preset threshold, step S207 is executed.

[0144] Among them, the detailed process of calculating the page health can be referred to step S103 in the above Figure 3 corresponding embodiment.

[0145] Step S207, terminate the release of the target version program and roll back to the historical version program.

[0146] Specifically, if the page health degree is not greater than a preset threshold, the release of the target version program is terminated, and 10% of the server sub-clusters are rolled back from the target version program to the historical version program.

[0147] Step S208, the intelligent continuous deployment system deploys the target version program to 50% of the formal environment.

[0148] Specifically, if the page health degree is greater than the preset threshold, the container platform is notified to deploy the target version program to 50% of the server sub-clusters of the full-scale servers. Of course, these 50% of the server sub-clusters include the 10% of the server sub-clusters mentioned above.

[0149] Step S209, 50% of the server sub-clusters report service data to the monitoring platform.

[0150] Step S210, the health degree calculation system in the intelligent continuous deployment system pulls the service data reported by 50% of the server sub-clusters from the monitoring platform.

[0151] Step S211, the health degree calculation system calculates the page health degree according to the service data. If the page health degree is greater than the preset threshold, step S212 is executed; if the page health degree is not greater than the preset threshold, step S207 is executed.

[0152] Step S212, the intelligent continuous deployment system deploys the target version program to 100% of the formal environment.

[0153] Specifically, if the page health degree is greater than the preset threshold, the container platform is notified to deploy the target version program to the full-scale servers.

[0154] At this point, the upgrade of the target version program is completed.

[0155] Step S213, the full-scale servers report service data to the monitoring platform.

[0156] This application uses the monitoring of the health degree in the gray release process of the Web front-end product. It can judge whether to continue to expand the gray scale or stop the release through the health degree index. When there is no abnormality, there is no need for manual intervention. When an abnormality occurs, manual intervention can be carried out in a timely manner, and the location where the abnormality occurs can be quickly located, without the need for manual monitoring of the gray scale situation. This greatly improves the release efficiency of the Web front-end product.

[0157] Further, please refer to Figure 8 , which is a schematic structural diagram of a program processing device provided by an embodiment of this application. As Figure 8 shown, the program processing device 1 can be applied to the above Figures 3 - 7For the program deployment device in the corresponding embodiment, the program processing device 1 may include a first acquisition module 11, a first deployment module 12, a second acquisition module 13, a status determination module 14, a policy determination module 15, and a second deployment module 16.

[0158] The first acquisition module 11 is used to acquire the target version program;

[0159] The first deployment module 12 is used to deploy the target version program to the first server sub-cluster in the server cluster, and the first server sub-cluster is a subset of the server cluster;

[0160] The second acquisition module 13 is used to acquire the first service data reported by the first server sub-cluster, and the first service data includes the operation behavior chain data of the user on the page corresponding to the target version program;

[0161] The status determination module 14 is used to determine the first page health status of the target version program according to the first service data;

[0162] The policy determination module 15 is used to determine the program deployment policy of the server cluster according to the first page health status;

[0163] The second deployment module 16 is used to deploy the program to the server cluster according to the program deployment policy.

[0164] The first page health status includes a page normal status or a page abnormal status;

[0165] The policy determination module 15 is specifically used for:

[0166] If the first page health status is the page normal status, then determine the continue deployment policy as the program deployment policy of the server cluster;

[0167] If the first page health status is the page abnormal status, then determine the rollback deployment policy as the program deployment policy of the server cluster; the continue deployment policy and the rollback deployment policy are different from each other.

[0168] The second acquisition module 13 is specifically used for:

[0169] Acquire the start timestamp when the target version program is deployed to the first server sub-cluster;

[0170] Determine the data reporting duration according to the start timestamp;

[0171] If the data reporting duration reaches the duration threshold, then acquire the service data associated with the target version program reported by the first server sub-cluster from the monitoring platform, and use the acquired service data as the first service data.

[0172] The first deployment module 12 is specifically configured to:

[0173] Submit the target version program to the container platform, so that the container platform selects the first server sub-cluster from the server cluster, and the container platform deploys the target version program to the first server sub-cluster.

[0174] Among them, the specific functional implementation manners of the first acquisition module 11, the first deployment module 12, the second acquisition module 13, the status determination module 14, the policy determination module 15, and the second deployment module 16 can refer to steps S101 - S104 in the corresponding embodiment above, which will not be elaborated here. Figure 3 Corresponding to steps S101 - S104 in the embodiment, details are not described herein again.

[0175] See again Figure 8 that the operation behavior chain data includes multiple user operation behaviors, and the first service data further includes the page access volume of each page where the user operation behavior is located;

[0176] The status determination module 14 may include: a superimposing unit 141, a first determination unit 142, and a second determination unit 143.

[0177] The superimposing unit 141 is configured to determine the number of behaviors of each user operation behavior according to the operation behavior chain data, obtain the first weight of each user operation behavior among the multiple user operation behaviors, and determine the behavior type of each user operation behavior;

[0178] The first determination unit 142 is configured to determine the behavior health degree of each user operation behavior according to the first weight of each user operation behavior, the behavior type of each user operation behavior, the number of behaviors of each user operation behavior, and the page access volume;

[0179] The superimposing unit 141 is further configured to superimpose the behavior health degrees of all user operation behaviors into a page health degree;

[0180] The second determination unit 143 is configured to determine the first page health status of the target version program according to the page health degree.

[0181] The first page health status includes a page normal status or a page abnormal status;

[0182] The second determination unit 143 is specifically configured to:

[0183] If the page health degree is greater than the health degree threshold, determine that the first page health status of the target version program is the page normal status;

[0184] If the page health degree is not greater than the health degree threshold, determine that the first page health state of the target version program is the page abnormal state.

[0185] The first determination unit 142 is specifically configured to:

[0186] Determine the operation rate of the user operation behavior according to the number of user operation behaviors and the page access volume of the page where the user operation behavior is located;

[0187] Determine the user operation parent behavior that is forward adjacent to the user operation behavior on the operation behavior chain data, and determine the conversion rate of the user operation behavior according to the number of behaviors of the user operation behavior and the number of behaviors of the user operation parent behavior;

[0188] Determine the second weight of the user operation behavior according to the number of behaviors of the user operation behavior and the number of behaviors of all user operation behaviors on the page where the user operation behavior is located;

[0189] Obtain the historical operation rate and historical conversion rate of the user operation behavior; the historical operation rate and historical conversion rate are determined according to the historical service data reported by the server cluster where the historical version program has been deployed;

[0190] Determine the behavior health degree of the user operation behavior according to the first weight, behavior type, operation rate, conversion rate, second weight, historical operation rate and historical conversion rate of the user operation behavior.

[0191] The number of conversion rates of the user operation behavior is N, and N is a positive integer;

[0192] The program processing device 1 may include a first acquisition module 11, a first deployment module 12, a second acquisition module 13, a state determination module 14, a policy determination module 15, and a second deployment module 16; it may also include: an exception determination module 17.

[0193] The exception determination module 17 is configured to obtain N historical conversion rates, determine the abnormal conversion rate from the N conversion rates according to the N historical conversion rates and the N conversion rates, determine the abnormal operation behavior chain data corresponding to the abnormal conversion rate, determine the abnormal page according to the abnormal operation behavior chain data and the user operation behavior, and output the abnormal page.

[0194] Among them, the specific function implementation manners of the superposition unit 141, the first determination unit 142, the second determination unit 143, and the exception determination module 17 may refer to step S103 in the corresponding embodiment above, and will not be elaborated here. Figure 3 Corresponding to step S103 in the embodiment, it will not be elaborated here.

[0195] Please refer to Figure 8, the program deployment policy includes the continue deployment policy;

[0196] The second deployment module 16 may include: a continue deployment unit 161, an acquisition unit 162, and a first rollback unit 163.

[0197] The continue deployment unit 161 is configured to deploy the target version program to the second server sub-cluster in the server cluster according to the continue deployment policy; the second server sub-cluster is a subset of the server cluster, and the intersection between the first server sub-cluster and the second server sub-cluster is empty;

[0198] The acquisition unit 162 is configured to acquire second service data reported by the first server sub-cluster and the second server sub-cluster, and determine a second page health status of the target version program according to the second service data; the second page health status includes a page normal status or a page abnormal status;

[0199] The acquisition unit 162 is further configured to, if the second page health status is the page normal status, deploy the target version program to the servers in the server cluster other than the first server sub-cluster and the second server sub-cluster.

[0200] The first rollback unit 163 is configured to, if the second page health status is the page abnormal status, roll back the first server sub-cluster and the second server sub-cluster from the target version program to the historical version program.

[0201] Please refer to again Figure 8 , the program deployment policy includes the rollback deployment policy;

[0202] The second deployment module 16 may include: a second rollback unit 164.

[0203] The second rollback unit 164 is configured to roll back the first server sub-cluster from the target version program to the historical version program according to the rollback deployment policy.

[0204] It can be known that when the continue deployment unit 161, the acquisition unit 162, and the first rollback unit 163 execute corresponding steps, the second rollback unit 164 does not execute corresponding steps; when the second rollback unit 164 executes corresponding steps, the continue deployment unit 161, the acquisition unit 162, and the first rollback unit 163 do not execute corresponding steps.

[0205] Among them, the specific functional implementation manners of the continue deployment unit 161, the acquisition unit 162, the first rollback unit 163, and the second rollback unit 164 can refer to step S104 in the corresponding Figure 3 corresponding embodiment, and will not be elaborated here.

[0206] Further, please refer to Figure 9 , which is a schematic structural diagram of a computer device provided by an embodiment of the present invention. The above Figures 3 - 7 The program deployment device in the corresponding embodiment may be the computer device 1000. As Figure 9 shown, the computer device 1000 may include: a user interface 1002, a processor 1004, an encoder 1006, and a memory 1008. The signal receiver 1016 is configured to receive or transmit data via a cellular interface 1010, a WIFI interface 1012, ..., or an NFC interface 1014. The encoder 1006 encodes the received data into a data format processed by a computer. A computer program is stored in the memory 1008, and the processor 1004 is configured to execute the steps in any of the above method embodiments through the computer program. The memory 1008 may include a volatile memory (e.g., dynamic random access memory DRAM), and may further include a non-volatile memory (e.g., one-time programmable read-only memory OTPROM). In some instances, the memory 1008 may further include a memory remotely provided with respect to the processor 1004, and these remote memories may be connected to the computer device 1000 through a network. The user interface 1002 may include: a keyboard 1018 and a display 1020.

[0207] In Figure 9 the computer device 1000 shown, the processor 1004 may be configured to call the computer program stored in the memory 1008 to implement:

[0208] Obtain a target version program, and deploy the target version program to a first server sub-cluster in the server cluster, where the first server sub-cluster is a subset of the server cluster;

[0209] Obtain first service data reported by the first server sub-cluster, where the first service data includes operation behavior chain data of a user on a page corresponding to the target version program;

[0210] Determine a first page health status of the target version program according to the first service data;

[0211] Determine a program deployment strategy for the server cluster according to the first page health status, and perform program deployment on the server cluster according to the program deployment strategy.

[0212] In one embodiment, the operation behavior chain data includes multiple user operation behaviors, and the first service data further includes the page access volume of each page where the user operation behavior is located;

[0213] When the processor 1004 executes to determine the first page health status of the target version program according to the first service data, the following steps are specifically executed:

[0214] Determine the number of actions for each user operation behavior according to the operation behavior chain data;

[0215] Obtain the first weight of each user operation behavior among the multiple user operation behaviors, and determine the behavior type of each user operation behavior;

[0216] Determine the behavior health degree of each user operation behavior according to the first weight of each user operation behavior, the behavior type of each user operation behavior, the number of actions of each user operation behavior, and the page access volume;

[0217] Superimpose the behavior health degrees of all user operation behaviors as the page health degree, and determine the first page health status of the target version program according to the page health degree.

[0218] In one embodiment, the first page health status includes a page normal status or a page abnormal status;

[0219] When the processor 1004 executes to determine the first page health status of the target version program according to the page health degree, the following steps are specifically executed:

[0220] If the page health degree is greater than the health degree threshold, determine that the first page health status of the target version program is the page normal status;

[0221] If the page health degree is not greater than the health degree threshold, determine that the first page health status of the target version program is the page abnormal status.

[0222] In one embodiment, when the processor 1004 executes to determine the behavior health degree of each user operation behavior according to the first weight of each user operation behavior, the behavior type of each user operation behavior, the number of actions of each user operation behavior, and the page access volume, the following steps are specifically executed:

[0223] Determine the operation rate of the user operation behavior according to the number of actions of the user operation behavior and the page access volume of the page where the user operation behavior is located;

[0224] Determine the user operation parent behavior adjacent to the user operation behavior in the forward direction on the operation behavior chain data, and determine the conversion rate of the user operation behavior according to the number of actions of the user operation behavior and the number of actions of the user operation parent behavior;

[0225] Determine the second weight of the user operation behavior according to the number of behaviors of the user operation behavior and the number of all user operation behaviors on the page where the user operation behavior is located;

[0226] Obtain the historical operation rate and historical conversion rate of the user operation behavior; the historical operation rate and historical conversion rate are determined according to the historical business data reported by the server cluster where the historical version program has been deployed;

[0227] Determine the behavior health of the user operation behavior according to the first weight, behavior type, operation rate, conversion rate, second weight, historical operation rate and historical conversion rate of the user operation behavior.

[0228] In one embodiment, the number of conversion rates of the user operation behavior is N, and N is a positive integer;

[0229] The processor 1004 also executes the following steps:

[0230] Obtain N historical conversion rates, and determine the abnormal conversion rate from the N conversion rates according to the N historical conversion rates and the N conversion rates;

[0231] Determine the abnormal operation behavior chain data corresponding to the abnormal conversion rate, and determine the abnormal page according to the abnormal operation behavior chain data and the user operation behavior, and output the abnormal page.

[0232] In one embodiment, the first page health state includes a page normal state or a page abnormal state;

[0233] When the processor 1004 executes to determine the program deployment strategy of the server cluster according to the first page health state, it specifically executes the following steps:

[0234] If the first page health state is the page normal state, then determine the continue deployment strategy as the program deployment strategy of the server cluster;

[0235] If the first page health state is the page abnormal state, then determine the rollback deployment strategy as the program deployment strategy of the server cluster; the continue deployment strategy and the rollback deployment strategy are different from each other.

[0236] In one embodiment, the program deployment strategy includes the continue deployment strategy;

[0237] When the processor 1004 executes to perform program deployment on the server cluster according to the program deployment strategy, it specifically executes the following steps:

[0238] Deploy the target version program to the second server sub - cluster in the server cluster according to the continued deployment strategy; the second server sub - cluster is a subset of the server cluster, and the intersection between the first server sub - cluster and the second server sub - cluster is empty;

[0239] Obtain the second service data reported by the first server sub - cluster and the second server sub - cluster, and determine the second page health status of the target version program according to the second service data; the second page health status includes a page normal status or a page abnormal status;

[0240] If the second page health status is the page normal status, deploy the target version program to the servers in the server cluster other than the first server sub - cluster and the second server sub - cluster.

[0241] In one embodiment, the processor 1004 also performs the following steps:

[0242] If the second page health status is the page abnormal status, roll back the first server sub - cluster and the second server sub - cluster from the target version program to the historical version program.

[0243] In one embodiment, when the processor 1004 performs program deployment on the server cluster according to the program deployment strategy, it specifically performs the following steps:

[0244] Roll back the first server sub - cluster from the target version program to the historical version program according to the roll - back deployment strategy.

[0245] In one embodiment, when the processor 1004 performs obtaining the first service data reported by the first server sub - cluster, it specifically performs the following steps:

[0246] Obtain the start timestamp when the target version program is deployed to the first server sub - cluster;

[0247] Determine the data reporting duration according to the start timestamp;

[0248] If the data reporting duration reaches the duration threshold, obtain the service data associated with the target version program reported by the first server sub - cluster from the monitoring platform, and use the obtained service data as the first service data.

[0249] In one embodiment, when the processor 1004 performs deploying the target version program to the first server sub - cluster in the server cluster, it specifically performs the following steps:

[0250] The target version program is submitted to the container platform, so that the container platform selects the first server sub-cluster from the server cluster, and the container platform deploys the target version program to the first server sub-cluster.

[0251] It should be understood that the computer device 1000 described in the embodiment of the present invention can execute the above Figures 3 to 7 The description of the program processing method in the corresponding embodiment can also be performed as described above. Figure 8 The description of the program processing device 1 in the corresponding embodiment will not be repeated here. In addition, the description of the beneficial effects of the same method will not be repeated here either.

[0252] In addition, it should be pointed out that: the embodiment of the present invention further provides a computer storage medium, and the computer storage medium stores a computer program executed by the program processing device 1 mentioned above, and the computer program includes program instructions. When the processor executes the program instructions, it can execute the above-mentioned Figures 3 to 7 The description of the program processing method in the corresponding embodiment will not be repeated here. In addition, the description of the beneficial effects of the same method will not be repeated. For technical details not disclosed in the computer storage medium embodiment involved in the present invention, please refer to the description of the method embodiment of the present invention.

[0253] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device can perform the above Figures 3 to 7 The method in the corresponding embodiment will therefore not be described in detail here.

[0254] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0255] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A program processing method, characterized in that, Including: Obtain the target version program and deploy the target version program to the first server sub - cluster in the server cluster, where the first server sub - cluster is a subset of the server cluster; Obtain the first service data reported by the first server sub - cluster. The first service data includes the operation behavior chain data of users on the page corresponding to the target version program. The operation behavior chain data includes multiple user operation behaviors, and the first service data also includes the page access volume of each page where the user operation behavior is located; Determine the behavior quantity of each user operation behavior according to the operation behavior chain data; Obtain the first weight of each user operation behavior among the multiple user operation behaviors, and determine the behavior type of each user operation behavior; Determine the behavior health degree of each user operation behavior according to the first weight, behavior type, behavior quantity of each user operation behavior, and the page access volume; Overlay the behavior health degrees of all user operation behaviors as the page health degree, and determine the first page health state of the target version program according to the page health degree; Determine the program deployment strategy of the server cluster according to the first page health state, and deploy the program to the server cluster according to the program deployment strategy.

2. The method according to claim 1, wherein The first page health state includes a page normal state or a page abnormal state; The determining the first page health state of the target version program according to the page health degree includes: If the page health degree is greater than the health degree threshold, determine that the first page health state of the target version program is the page normal state; If the page health degree is not greater than the health degree threshold, determine that the first page health state of the target version program is the page abnormal state.

3. The method according to claim 1, wherein The determining the behavior health degree of each user operation behavior according to the first weight, behavior type, behavior quantity of each user operation behavior, and the page access volume includes: Determine the operation rate of the user operation behavior according to the quantity of the user operation behavior and the page access volume of the page where the user operation behavior is located; Determine the user operation parent behavior adjacent to the user operation behavior forward on the operation behavior chain data, and determine the conversion rate of the user operation behavior according to the behavior quantity of the user operation behavior and the behavior quantity of the user operation parent behavior; Determine the second weight of the user operation behavior according to the behavior quantity of the user operation behavior and the behavior quantity of all user operation behaviors on the page where the user operation behavior is located; Obtain the historical operation rate and historical conversion rate of the user operation behavior; the historical operation rate and historical conversion rate are determined according to the historical service data reported by the server cluster where the historical version program has been deployed; Determine the behavior health degree of the user operation behavior according to the first weight, behavior type, operation rate, conversion rate, second weight, historical operation rate, and historical conversion rate of the user operation behavior.

4. The method according to claim 3, wherein The number of conversion rates of the user operation behavior is N, and N is a positive integer; The method further includes: Obtain N historical conversion rates, and determine abnormal conversion rates from the N conversion rates according to the N historical conversion rates and the N conversion rates; Determine the abnormal operation behavior chain data corresponding to the abnormal conversion rates, determine the abnormal page according to the abnormal operation behavior chain data and the user operation behavior, and output the abnormal page.

5. The method according to claim 1, wherein The health status of the first page includes a page normal status or a page abnormal status; The determining the program deployment strategy of the server cluster according to the health status of the first page includes: If the health status of the first page is the page normal status, determine the continue deployment strategy as the program deployment strategy of the server cluster; If the health status of the first page is the page abnormal status, determine the rollback deployment strategy as the program deployment strategy of the server cluster; the continue deployment strategy and the rollback deployment strategy are different from each other.

6. The method according to claim 5, wherein The program deployment strategy includes the continue deployment strategy; The performing program deployment on the server cluster according to the program deployment strategy includes: According to the continue deployment strategy, deploy the target version program to the second server sub-cluster in the server cluster; the second server sub-cluster is a subset of the server cluster, and the intersection between the first server sub-cluster and the second server sub-cluster is empty; Obtain the second service data reported by the first server sub-cluster and the second server sub-cluster, and determine the health status of the second page of the target version program according to the second service data; the health status of the second page includes a page normal status or a page abnormal status; If the health status of the second page is the page normal status, deploy the target version program to the servers in the server cluster other than the first server sub-cluster and the second server sub-cluster.

7. The method according to claim 6, wherein It also includes: If the health status of the second page is the page abnormal status, roll back the first server sub-cluster and the second server sub-cluster from the target version program to the historical version program.

8. The method according to claim 5, wherein The program deployment strategy includes the rollback deployment strategy; The performing program deployment on the server cluster according to the program deployment strategy includes: According to the rollback deployment strategy, roll back the first server sub-cluster from the target version program to the historical version program.

9. The method according to claim 1, wherein The obtaining the first service data reported by the first server sub-cluster includes: Obtain the start timestamp when the target version program is deployed to the first server sub-cluster; According to the start timestamp, determine the data reporting duration; If the data reporting duration reaches the duration threshold, obtain the service data associated with the target version program reported by the first server sub-cluster from the monitoring platform, and use the obtained service data as the first service data.

10. The method according to claim 1, characterized in that The deploying the target version program to the first server sub-cluster in the server cluster includes: Submit the target version program to the container platform, so that the container platform selects the first server sub-cluster from the server cluster, and the container platform deploys the target version program to the first server sub-cluster.

11. A program processing device, characterized in that, It includes: The first acquisition module is used to acquire the target version program; The first deployment module is used to deploy the target version program to the first server sub - cluster in the server cluster, and the first server sub - cluster is a subset of the server cluster; The second acquisition module is used to acquire the first service data reported by the first server sub - cluster. The first service data includes operation behavior chain data of users on the page corresponding to the target version program. The operation behavior chain data includes multiple user operation behaviors, and the first service data further includes the page access volume of each page where the user operation behavior is located; The status determination module is used to determine the behavior quantity of each user operation behavior according to the operation behavior chain data; Obtain the first weight of each user operation behavior among the multiple user operation behaviors, and determine the behavior type of each user operation behavior; Determine the behavior health degree of each user operation behavior according to the first weight, behavior type, behavior quantity of each user operation behavior, and the page access volume; superimpose the behavior health degrees of all user operation behaviors as the page health degree, and determine the first page health status of the target version program according to the page health degree; The policy determination module is used to determine the program deployment policy of the server cluster according to the first page health status; The second deployment module is used to deploy the program to the server cluster according to the program deployment policy.

12. A computer device, characterized in that, It includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 - 10.

13. A computer storage medium, characterized in that, The computer storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by the processor, the method according to any one of claims 1 - 10 is executed.

14. A computer program product, which includes computer instructions. The computer instructions are stored in a computer - readable storage medium. The processor of the computer device reads and executes the computer instructions, so that the computer device executes the method according to any one of claims 1 - 10.

Images