A cloud-based website performance monitoring method and system

Abstract

The application discloses a cloud-based website performance monitoring method and system, and relates to the technical field of website performance monitoring.The method comprises the following steps: encapsulating a monitoring node as a lightweight container, and selecting an optimal running node for the monitoring node in a hybrid cloud node pool; sending a request to a website by a self-adaptive simulation client of the monitoring node, and capturing multi-dimensional response data of the target website; quantifying each running state of the target website in real time based on the captured multi-dimensional response data; calculating a performance score according to the quantification results of each running state of the target website; and visually displaying the performance score of each website in time sequence, and sending an alarm notification when the performance score is lower than a threshold value.The flexible deployment of the monitoring node can be realized by increasing or decreasing cloud nodes; the method can be adapted to increasingly updated Web technologies; multi-dimensional response data monitoring is supported, and more accurate health detection information can be generated.

Description

Technical Field

[0001] This invention relates to the field of website performance monitoring technology, and in particular to a cloud-based website performance monitoring method and system. Background Technology

[0002] With the rapid development of internet technology, website system architectures are becoming increasingly complex, exhibiting characteristics such as distributed services, multi-CDN acceleration, and dynamic content loading. Against this backdrop, website performance monitoring technology has become a crucial element in ensuring user experience and maintaining service stability. However, existing technical solutions mostly adopt a localized deployment model, relying on monitoring nodes in fixed locations (such as internal enterprise servers), resulting in a monitoring scope limited by the geographical distribution of preset nodes. Current methods also primarily focus on collecting basic performance indicators (such as response time and HTTP status codes), lacking adaptability to new web technologies. To address these technical shortcomings, there is an urgent need to develop a new cloud-based monitoring method to achieve dynamically scalable monitoring node deployment, adapt to ever-evolving web technologies, and thus build a more accurate and efficient website performance monitoring system. Summary of the Invention

[0003] This invention provides a cloud-based website performance monitoring method, comprising:

[0004] Step 1: Encapsulate the monitoring node into a lightweight container and select the optimal running node for it in the hybrid cloud node pool;

[0005] Step 2: By adaptively simulating the client to send requests to the website through monitoring nodes, capture the multi-dimensional response data of the target website;

[0006] Step 3: Quantify the various operational statuses of the target website in real time based on the captured multidimensional response data;

[0007] Step 4: Calculate the performance score based on the quantitative results of the target website's various operational statuses;

[0008] Step 5: Visualize the performance scores of each website over time and issue an alarm notification when the performance score is below the threshold.

[0009] The cloud-based website performance monitoring method described above encapsulates the monitoring nodes as lightweight containers and selects the optimal running node for them from a hybrid cloud node pool. Specifically, it consists of the following sub-steps:

[0010] A base image is built based on the latest stable version of Chromium, and the monitoring scripts and configuration files of the target website are packaged into the image to generate a well-packaged lightweight container.

[0011] By deploying nodes from multiple cloud vendors and edge computing nodes, a hybrid cloud node pool is formed;

[0012] Based on user geographic distribution, target website CDN configuration, and historical monitoring data, the optimal running node is dynamically selected for the packaged lightweight container.

[0013] The cloud-based website performance monitoring method described above involves capturing multi-dimensional response data from the target website by adaptively simulating client requests to the website through monitoring nodes. This method comprises the following sub-steps:

[0014] Build a browser environment based on the web technology stack used by the target website;

[0015] Dynamically adjust request parameters and use the constructed browser environment to send requests to the target website;

[0016] Extract and record the multidimensional response data returned by the target website into a separate data table.

[0017] The cloud-based website performance monitoring method described above quantifies various operational statuses of the target website in real time based on captured multidimensional response data, and specifically consists of the following sub-steps:

[0018] The captured multidimensional response data is standardized.

[0019] The latent variables of the multidimensional response data are extracted using the pre-model, and the multidimensional response data is reconstructed based on the latent variables;

[0020] The reconstruction error, potential spatial characteristics, and dynamic adjustment factors of the multidimensional response data are used to quantify the various operational statuses of the target website.

[0021] The present invention also provides a cloud-based website performance monitoring system, including: a monitoring node deployment module, a request simulation module, a running status quantification module, a performance score generation module, and a performance score visualization module;

[0022] The monitoring node deployment module is used to encapsulate monitoring nodes as lightweight containers and select the optimal running node for them in the hybrid cloud node pool;

[0023] The request simulation module is used to adaptively simulate the client sending requests to the website through monitoring nodes, and capture multi-dimensional response data from the target website.

[0024] The operational status quantification module is used to quantify the various operational statuses of the target website in real time based on the captured multidimensional response data;

[0025] The performance score generation module is used to calculate the performance score based on the quantitative results of various operational statuses of the target website.

[0026] The performance rating visualization module is used to visually display the performance rating of each website over time and to issue an alarm notification when the performance rating falls below a threshold.

[0027] The beneficial effects achieved by this invention are as follows: the monitoring nodes can be flexibly deployed by adding or removing cloud nodes; it can adapt to the ever-updating Web technologies; it supports multi-dimensional response data monitoring and can generate more accurate health detection information. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0029] Figure 1 This is a flowchart of a cloud-based website performance monitoring method provided in Embodiment 1 of this application;

[0030] Figure 2 This is a schematic diagram of a cloud-based website performance monitoring system provided in Embodiment 2 of this application. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Example 1

[0033] like Figure 1 As shown, Embodiment 1 of this application provides a cloud-based website performance monitoring method, including:

[0034] Step S10: Encapsulate the monitoring node into a lightweight container and select the optimal running node for it in the hybrid cloud node pool;

[0035] Encapsulating monitoring nodes into lightweight containers can improve resource utilization, while deploying them to cloud nodes enables dynamic scaling and flexible deployment of monitoring nodes. Specifically:

[0036] Step S11: Build a base image based on the latest stable version of Chromium, and package the target website's monitoring scripts and configuration files into the image to generate a packaged lightweight container;

[0037] The monitoring scripts include script files for initializing monitoring nodes, such as importing dependency libraries and loading configurations. The configuration files include monitoring target configurations, multi-region configurations, and compliance configurations. It is recommended to use CI / CD pipelines to automatically build images to speed up system initialization time.

[0038] Step S12: Form a hybrid cloud node pool by deploying nodes from multiple cloud vendors and edge computing nodes;

[0039] This hybrid cloud node pool allows monitoring nodes to be deployed close to users, reducing network latency, while leveraging the redundancy of multi-cloud resources to improve system availability.

[0040] Step S13: Based on the user's geographic distribution, the target website's CDN configuration, and historical monitoring data, dynamically select the optimal running node for the packaged lightweight container;

[0041] Every 5 seconds, the suitability index of each cloud node for the current container is calculated. The cloud node with the highest suitability index is dynamically selected to run the current container. If the load of the node with the highest index exceeds 80%, the next highest-index cloud node is selected. The cloud node N is then calculated. k The formula for the applicable index Z of the current container is expressed as:

[0042]

[0043] Where ω1, ω2, and ω3 are adjustable weights, Latency(u,N) k This indicates the distance from user u to cloud node N obtained through IP speed testing. k Round trip time, CDNTier(N) k ) represents cloud node N k In the target website CDN, role assignments are defined (e.g., edge node = 1, regional relay = 2, origin server = 3), max(CDNTier) represents the maximum role assignment for a cloud node, α and β are adjustable resolution coefficients, and AvgRTT(N k ) represents cloud node N k The average response time for requests to the target website in a historical window (e.g., the past hour), ER(N) k ) represents cloud node N k The probability of the target website returning a server error in the history window, RTT max ER max These represent the maximum response time and the maximum server error rate across all cloud nodes.

[0044] Step S20: By adaptively simulating the client to send requests to the website through the monitoring node, capture the multi-dimensional response data of the target website;

[0045] Adaptive simulation of a client sending requests to a website refers to the monitoring node dynamically constructing a browser environment based on the website's technology stack to simulate a client sending requests to the website, in order to adapt to the ever-evolving web technologies. This process involves the following sub-steps:

[0046] Step S21: Build a browser environment based on the web technology stack used by the target website;

[0047] The browser kernel and plugin components are separated, and a daily automated pipeline is built to synchronize with upstream Chromium updates, ensuring that the kernel version lags by no more than 72 hours. Based on the web technology stack used by the target website, the appropriate plugin components are loaded on demand. For example, the web technology stack is identified by parsing the web page HTML, detecting front-end framework feature tags, extracting JS / CSS file URLs, and detecting filename hashes or path patterns. Then, the required plugin components for the corresponding technology stack are retrieved from the library and loaded into the current browser environment.

[0048] Step S22: Dynamically adjust the request parameters and send a request to the target website using the constructed browser environment;

[0049] The request parameters that need to be dynamically adjusted include: user region, device fingerprint, and request frequency. Different user regions can be simulated by switching the exit IP through a proxy pool or by using the GeoIP database to match the target website's CDN nodes. Different device fingerprints such as PC, mobile, and crawler can be simulated by modifying the User-Agent header. The request frequency can be dynamically adjusted based on the token bucket algorithm to avoid triggering the website's anti-crawling mechanism. After configuration, the built browser environment can be used to send requests to the target website.

[0050] Step S23: Extract and record the multidimensional response data returned by the target website into a separate data table;

[0051] Multidimensional response data includes data from three dimensions: basic metrics, front-end metrics, and API monitoring metrics. Basic metrics include fundamental performance indicators such as network latency (RTT), TCP connection time, TLS handshake time, and server response time (TTFB). Front-end metrics refer to key front-end performance indicators captured by monitoring nodes during the page loading process by executing JavaScript code, including first content rendering (FCP), maximum content rendering (LCP), and cumulative layout offset (CLS). API monitoring metrics refer to monitoring nodes sending simulated requests to the API interfaces provided by the target website and recording the interface response time, success rate, and the correctness of the returned data.

[0052] Step S30: Based on the captured multidimensional response data, quantify the various operational statuses of the target website in real time, specifically divided into the following sub-steps:

[0053] Step S31: Standardize the captured multidimensional response data;

[0054] If the response data has a high dimension (e.g., more than 10 dimensions), dimensionality reduction can be achieved through PCA or t-SNE, preserving 95% of the variance.

[0055] Step S32: Use the pre-model to extract latent variables from the multidimensional response data, and reconstruct the multidimensional response data based on the latent variables;

[0056] The standardized multidimensional response data are input into the pre-model to obtain the reconstructed dataset. The pre-model consists of three parts: an encoder, a sampler, and a decoder. The encoder maps the input data X to the latent space and contains two neural networks f. μ and f σ (Both are fully connected layers with ReLU activation), used to return the mean μ and log-variance logσ of the latent variables generated after X is mapped to the latent space, respectively. 2 The sampler is used to apply the formula: Complete the sampling of latent variables, where b is the returned latent variable, and ∈ represents standard normal distribution noise. This is an element-wise product; the decoder is used to reconstruct the latent variable b returned by the sampler into the output. The decoder neural network consists of a fully connected layer with Sigmoid activation.

[0057] Step S33: Quantify the various operational statuses of the target website by integrating the reconstruction error, potential spatial characteristics, and dynamic adjustment factors of the multidimensional response data;

[0058] The quantification formula for the i-th running state of the target website at time t is expressed as:

[0059]

[0060] Where H it This represents the quantification result of the i-th running state of the target website at time t. For adjustable weights, It is the reconstruction error term. It is a latent spatial characteristic term, (1+φ·e -λt )and It is a dynamically adjusted factor, X it Let be the value of the i-th dimension of the response data at time t. For X it The reconstructed value, b it Let μ be the latent variable sampled from the i-th dimension response data by the pre-model at time t. b , Let represent the mean and covariance matrix of all latent variables generated by the i-th dimension response data in the latent space, respectively; φ is the decay factor; λ is the decay rate; and γ is the sensitivity coefficient for special events. This flag indicates whether a special event (such as a promotion on an e-commerce website) exists at time t; if it does, the value is 1, otherwise it is 0.

[0061] Step S40: Calculate the performance score based on the quantitative results of the target website's various operational statuses;

[0062] Performance rating S t The formula used to reflect the overall health status of the website at time t is as follows:

[0063] in H represents the emphasis coefficient of the target website on the i-th operational status. it H represents the quantification result of the i-th running status of the target website at time t. i0 τ represents the quantitative result of the i-th operating state of the target website under standard conditions. i Let i be the maximum tolerance of the target website for the i-th running state, where i ranges from 1 to n, and n is the total number of running states.

[0064] Step S50: Visualize the performance scores of each website over time and issue an alarm notification when the performance score is lower than the threshold;

[0065] After obtaining the performance scores of each website over time, they need to be displayed visually so that operations and maintenance personnel can intuitively obtain the monitoring results. When the performance score falls below the threshold, an alert notification should be promptly sent to the corresponding operations and maintenance personnel. Specifically:

[0066] Step S51: Generate a single-website visual monitoring table with time as the horizontal axis and the target website's performance score as the vertical axis;

[0067] Real-time plotting is performed using Python libraries such as Matplotlib and Plotly. Appropriate chart types, such as line charts and area charts, are selected according to requirements to show how performance scores change over time. Then, the latest score data is periodically retrieved through scheduled tasks or event-driven mechanisms to update the visualization monitoring table to ensure the real-time nature of the visualization.

[0068] Step S52: Synchronize the monitoring timestamps of multiple websites and generate a multi-website visual monitoring table;

[0069] If multiple websites are selected when generating the visualization monitoring table, you first need to synchronize the monitoring timestamps of each website and then plot them into the same table.

[0070] Step S53: Add a threshold line to the visual monitoring table. When the performance score of a website is detected to be lower than the threshold, an alarm mechanism is triggered.

[0071] Add a threshold line to visually indicate whether the performance score is below a preset threshold. The threshold is reasonably set based on business needs and historical performance score data. Each time the chart is updated, the performance score data is checked simultaneously. When a website's performance score is found to be below the threshold, an alarm mechanism is triggered. The alarm notification is sent via email, SMS, pop-up window, etc., and should include key information such as website name, timestamp, and performance score to quickly locate the problem.

[0072] Example 2

[0073] like Figure 2 As shown, Embodiment 2 of this application provides a cloud-based website performance monitoring system, including: a monitoring node deployment module 21, a request simulation module 22, a running status quantification module 23, a performance score generation module 24, and a performance score visualization module 25;

[0074] The monitoring node deployment module 21 is used to encapsulate the monitoring node into a lightweight container and select the optimal running node for it in the hybrid cloud node pool; specifically, it includes: a container encapsulation submodule, a hybrid cloud node pool deployment submodule, and a running node dynamic selection submodule.

[0075] 1. Container encapsulation submodule, used to build a base image based on the latest stable version of Chromium, and package the target website's monitoring scripts and configuration files into the image to generate a well-encapsulated lightweight container;

[0076] The monitoring scripts include script files for initializing monitoring nodes, such as importing dependency libraries and loading configurations. The configuration files include monitoring target configurations, multi-region configurations, and compliance configurations. It is recommended to use CI / CD pipelines to automatically build images to speed up system initialization time.

[0077] 2. Hybrid Cloud Node Pool Deployment Submodule: This module is used to form a hybrid cloud node pool by deploying nodes from multiple cloud vendors and edge computing nodes.

[0078] This hybrid cloud node pool allows monitoring nodes to be deployed close to users, reducing network latency, while leveraging the redundancy of multi-cloud resources to improve system availability.

[0079] 3. Dynamic Node Selection Submodule: This module dynamically selects the optimal running node for the packaged lightweight container based on user geographic distribution, target website CDN configuration, and historical monitoring data.

[0080] Every 5 seconds, the suitability index of each cloud node for the current container is calculated. The cloud node with the highest suitability index is dynamically selected to run the current container. If the load of the node with the highest index exceeds 80%, the next highest-index cloud node is selected. The cloud node N is then calculated. k The formula for the applicable index Z of the current container is expressed as:

[0081]

[0082] Where ω1, ω2, and ω3 are adjustable weights, Latency(u,N) k This indicates the distance from user u to cloud node N obtained through IP speed testing. k Round trip time, CDNTier(N) k ) represents cloud node N k In the target website CDN, role assignments are defined (e.g., edge node = 1, regional relay = 2, origin server = 3), max(CDNTier) represents the maximum role assignment for a cloud node, α and β are adjustable resolution coefficients, and AvgRTT(N k ) represents cloud node N k The average response time for requests to the target website in a historical window (e.g., the past hour), ER(N) k ) represents cloud node N k The probability of the target website returning a server error in the history window, RTT max ER max These represent the maximum response time and the maximum server error rate across all cloud nodes.

[0083] The request simulation module 22 is used to adaptively simulate the client sending requests to the website through the monitoring node and capture the multi-dimensional response data of the target website; specifically, it includes: a browser environment construction submodule, a request parameter adjustment submodule, and a multi-dimensional response data capture submodule.

[0084] 1. Browser environment building submodule, used to build the browser environment based on the Web technology stack used by the target website;

[0085] The browser kernel and plugin components are separated, and a daily automated pipeline is built to synchronize with upstream Chromium updates, ensuring that the kernel version lags by no more than 72 hours. Based on the web technology stack used by the target website, the appropriate plugin components are loaded on demand. For example, the web technology stack is identified by parsing the web page HTML, detecting front-end framework feature tags, extracting JS / CSS file URLs, and detecting filename hashes or path patterns. Then, the required plugin components for the corresponding technology stack are retrieved from the library and loaded into the current browser environment.

[0086] 2. The request parameter adjustment submodule is used to dynamically adjust request parameters and send requests to the target website using the built browser environment;

[0087] The request parameters that need to be dynamically adjusted include: user region, device fingerprint, and request frequency. Different user regions can be simulated by switching the exit IP through a proxy pool or by using the GeoIP database to match the target website's CDN nodes. Different device fingerprints such as PC, mobile, and crawler can be simulated by modifying the User-Agent header. The request frequency can be dynamically adjusted based on the token bucket algorithm to avoid triggering the website's anti-crawling mechanism. After configuration, the built browser environment can be used to send requests to the target website.

[0088] 3. Multidimensional response data capture submodule, used to extract and record the multidimensional response data returned by the target website into a separate data table;

[0089] Multidimensional response data includes data from three dimensions: basic metrics, front-end metrics, and API monitoring metrics. Basic metrics include fundamental performance indicators such as network latency (RTT), TCP connection time, TLS handshake time, and server response time (TTFB). Front-end metrics refer to key front-end performance indicators captured by monitoring nodes during the page loading process by executing JavaScript code, including first content rendering (FCP), maximum content rendering (LCP), and cumulative layout offset (CLS). API monitoring metrics refer to monitoring nodes sending simulated requests to the API interfaces provided by the target website and recording the interface response time, success rate, and the correctness of the returned data.

[0090] The operation status quantification module 23 is used to quantify the various operation statuses of the target website in real time based on the captured multidimensional response data. Specifically, it includes: a multidimensional response data preprocessing submodule, a multidimensional response data reconstruction submodule, and an operation status calculation submodule.

[0091] 1. Multidimensional response data preprocessing submodule, used to standardize the captured multidimensional response data;

[0092] If the response data has a high dimension (e.g., more than 10 dimensions), dimensionality reduction can be achieved through PCA or t-SNE, preserving 95% of the variance.

[0093] 2. The multidimensional response data reconstruction submodule uses the pre-model to extract the latent variables of the multidimensional response data and reconstructs the multidimensional response data based on the latent variables;

[0094] The standardized multidimensional response data are input into the pre-model to obtain the reconstructed dataset. The pre-model consists of three parts: an encoder, a sampler, and a decoder. The encoder maps the input data X to the latent space and contains two neural networks f. μ and f σ(Both are fully connected layers with ReLU activation), used to return the mean μ and log-variance logσ of the latent variables generated after X is mapped to the latent space, respectively. 2 The sampler is used to apply the formula: Complete the sampling of latent variables, where b is the returned latent variable, and ∈ represents standard normal distribution noise. This is an element-wise product; the decoder is used to reconstruct the latent variable b returned by the sampler into the output. The decoder neural network consists of a fully connected layer with Sigmoid activation.

[0095] 3. The comprehensive quantification submodule is used to quantify the various operational statuses of the target website by comprehensively considering the reconstruction error, potential spatial characteristics, and dynamic adjustment factors of multidimensional response data.

[0096] The quantification formula for the i-th running state of the target website at time t is expressed as:

[0097]

[0098] Where H it This represents the quantification result of the i-th running state of the target website at time t. For adjustable weights, It is the reconstruction error term. It is a latent spatial characteristic term, (1+φ·e -λt )and It is a dynamically adjusted factor, X it Let be the value of the i-th dimension of the response data at time t. For X it The reconstructed value, b it Let μ be the latent variable sampled from the i-th dimension response data by the pre-model at time t. b , Let represent the mean and covariance matrix of all latent variables generated by the i-th dimension response data in the latent space, respectively; φ is the decay factor; λ is the decay rate; and γ is the sensitivity coefficient for special events. This flag indicates whether a special event (such as a promotion on an e-commerce website) exists at time t; if it does, the value is 1, otherwise it is 0.

[0099] The performance score generation module 24 is used to calculate the performance score based on the quantitative results of various operating states of the target website;

[0100] Performance rating S t The formula used to reflect the overall health status of the website at time t is as follows:

[0101] in H represents the emphasis coefficient of the target website on the i-th operational status. itH represents the quantification result of the i-th running status of the target website at time t. i0 τ represents the quantitative result of the i-th operating state of the target website under standard conditions. i Let i be the maximum tolerance of the target website for the i-th running state, where i ranges from 1 to n, and n is the total number of running states.

[0102] The performance rating visualization module 25 is used to visually display the performance rating of each website over time and to issue an alarm notification when the performance rating is lower than the threshold; specifically, it includes: a visualization monitoring table generation submodule and an alarm triggering submodule.

[0103] 1. The Visual Monitoring Table Generation Submodule is used to generate visual monitoring tables for single or multiple websites with time as the horizontal axis and the performance score of the target website as the vertical axis.

[0104] Real-time plotting is performed using Python libraries such as Matplotlib and Plotly. Appropriate chart types, such as line charts and area charts, are selected based on requirements to display performance scores changing over time. Then, a scheduled task or event-driven mechanism is used to periodically retrieve the latest score data and update the visualization monitoring table to ensure real-time visualization. If multiple websites are selected when generating the visualization monitoring table, the monitoring timestamps of each website must first be synchronized before plotting them into the same table.

[0105] 2. Alarm Trigger Submodule: This module adds threshold lines to the visual monitoring table. When a website's performance score is detected to be below the threshold, an alarm mechanism is triggered.

[0106] Add a threshold line to visually indicate whether the performance score is below a preset threshold. The threshold is reasonably set based on business needs and historical performance score data. Each time the chart is updated, the performance score data is checked simultaneously. When a website's performance score is found to be below the threshold, an alarm mechanism is triggered. The alarm notification is sent via email, SMS, pop-up window, etc., and should include key information such as website name, timestamp, and performance score to quickly locate the problem.

[0107] Corresponding to the above embodiments, the present invention provides a computer storage medium, including: at least one memory and at least one processor;

[0108] The memory is used to store one or more program instructions;

[0109] A processor is used to run one or more program instructions to execute a cloud-based website performance monitoring method.

[0110] Corresponding to the above embodiments, this embodiment of the invention provides a computer-readable storage medium containing one or more program instructions, which are executed by a processor to provide a cloud-based website performance monitoring method.

[0111] The embodiments disclosed in this invention provide a computer-readable storage medium storing computer program instructions, which, when executed on a computer, cause the computer to perform the aforementioned cloud-based website performance monitoring method.

[0112] In this embodiment of the invention, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0113] The various methods, steps, and logic diagrams disclosed in the embodiments of this invention can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the above methods.

[0114] The storage medium can be memory, such as volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0115] Among them, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0116] Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0117] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0118] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in this invention can be implemented using a combination of hardware and software. When applied as software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0119] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. A cloud-based website performance monitoring method, characterized in that, Comprise: Step1, encapsulate the monitoring node into a lightweight container, and select the optimal running node for it in the hybrid cloud node pool, which is divided into the following substeps: Build a base image based on the latest stable version of Chromium, and package the monitoring script and configuration file of the target website into the image to generate a packaged lightweight container; Form a hybrid cloud node pool by deploying multi-cloud vendor nodes and edge computing nodes; According to the user's geographical distribution, the target website's CDN configuration and historical monitoring data, dynamically select the optimal running node for the packaged lightweight container; Every other preset period, calculate the suitability index of each cloud node for the current container, dynamically select the cloud node with the highest suitability index to run the current container, if the highest index node load exceeds the threshold, then delay to the second highest cloud node, calculate the suitability index of each cloud node for the current container k The formula of the suitability index Z of the current container is: , wherein is the adjustable weight, represents the round-trip time of user u to cloud node obtained by IP speed test, represents the role assignment of cloud node in the target website CDN, represents the maximum role assignment of cloud node, is the adjustable resolution coefficient, represents the average response time of cloud node requesting the target website in the history window, represents the probability of cloud node returning server error when requesting the target website in the history window, respectively the maximum response time and the maximum server error rate in all cloud nodes; Step2, send requests to the website through the monitoring node adaptive simulation client to capture the multi-dimensional response data of the target website; Step3, quantify the running status of the target website in real time based on the captured multi-dimensional response data; Step4, calculate the performance score according to the quantitative results of the running status of the target website; Step5, visualize the performance score of each website in time sequence, and send an alarm notification when the performance score is lower than the threshold.

2. The cloud-based website performance monitoring method of claim 1, wherein, Send requests to the website through the monitoring node adaptive simulation client to capture the multi-dimensional response data of the target website, which is divided into the following substeps: Build a browser environment according to the Web technology stack used by the target website; Dynamically adjust the request parameters and send requests to the target website using the built browser environment; Extract and record the multi-dimensional response data returned by the target website to a separate data table.

3. The cloud-based website performance monitoring method of claim 1, wherein, Quantify the running status of the target website in real time based on the captured multi-dimensional response data, which is divided into the following substeps: Standardize the captured multi-dimensional response data; Use the pre-model to extract the latent variables of the multi-dimensional response data, and reconstruct the multi-dimensional response data according to the latent variables; Quantify the running status of the target website by combining the reconstruction error, latent space features and dynamic adjustment factor of the multi-dimensional response data.

4. The cloud-based website performance monitoring method of claim 1, wherein, Visualize the performance score of each website in time sequence, and send an alarm notification when the performance score is lower than the threshold, which is divided into the following substeps: Take time as the horizontal coordinate and the performance score of the target website as the vertical coordinate to generate a single website visualization monitoring table; Synchronize the monitoring timestamps of multiple websites to generate a multi-website visualization monitoring table; Add a threshold line to the visualization monitoring table, and trigger the alarm mechanism when the performance score of a website is detected to be lower than the threshold.

5. A cloud-based website performance monitoring system, characterized by, A cloud-based website performance monitoring method as claimed in any one of claims 1-4, comprising: a monitoring node deployment module, a request simulation module, a running status quantification module, a performance score generation module, and a performance score visualization module; The monitoring node deployment module is used to encapsulate the monitoring node into a lightweight container and select the optimal running node for it in the hybrid cloud node pool; The request simulation module is used to send requests to the website through the monitoring node adaptive simulation client to capture the multi-dimensional response data of the target website; The running status quantification module is used to quantify the running status of the target website in real time based on the captured multi-dimensional response data; The performance score generation module is used to calculate the performance score according to the quantitative results of the running status of the target website; The performance score visualization module is configured to visualize the performance scores of each website over time and to send an alert notification when the performance score is below a threshold value.

6. The cloud-based website performance monitoring system of claim 5, wherein, The monitoring node deployment module specifically comprises a container packaging submodule, a hybrid cloud point pool deployment submodule, and a running node dynamic selection submodule. The container packaging submodule is configured to build a base image based on the latest stable version of Chromium, and to package the monitoring script and configuration file of the target website into the image to generate a packaged lightweight container. The hybrid cloud point pool deployment submodule is configured to form a hybrid cloud node pool by deploying multi-cloud vendor nodes and edge computing nodes. The running node dynamic selection submodule is configured to dynamically select the optimal running node for the packaged lightweight container based on the geographical distribution of users, the CDN configuration of the target website, and historical monitoring data.

7. The cloud-based website performance monitoring system of claim 5, wherein, The request simulation module specifically comprises a browser environment construction submodule, a request parameter adjustment submodule, and a multi-dimensional response data capture submodule. The browser environment construction submodule is configured to construct a browser environment based on the Web technology stack used by the target website. The request parameter adjustment submodule is configured to dynamically adjust the request parameters and send a request to the target website using the constructed browser environment. The multi-dimensional response data capture submodule is configured to extract and record the multi-dimensional response data returned by the target website into a separate data table.

8. The cloud-based website performance monitoring system of claim 5, wherein, The running state quantification module specifically comprises a multi-dimensional response data preprocessing submodule, a multi-dimensional response data reconstruction submodule, and a running state calculation submodule. The multi-dimensional response data preprocessing submodule is configured to standardize the captured multi-dimensional response data. The multi-dimensional response data reconstruction submodule is configured to extract the latent variables of the multi-dimensional response data using a preposition model, and to reconstruct the multi-dimensional response data based on the latent variables. The comprehensive quantification submodule is configured to quantify the running states of the target website based on the reconstruction error, latent space features, and dynamic adjustment factors of the multi-dimensional response data.

9. A computer storage medium, comprising, The system comprises: at least one memory and at least one processor; the memory is configured to store one or more program instructions; the processor is configured to run the one or more program instructions to execute the cloud-based website performance monitoring method according to any one of claims 1-4.

Images