An intelligent operation and maintenance system, method and device for micro-service application

Abstract

The application belongs to the technical field of system operation and maintenance, and discloses an intelligent operation and maintenance system, method, device and terminal for micro-service application, which comprises the following steps: an infrastructure layer manages hardware resources through containerization technology; an application layer develops applications by using micro-service technology; a management layer collects data of the infrastructure layer and the application layer; meanwhile, the management layer is used for registering, scheduling and operation and maintenance management of application instances; the infrastructure layer, the application layer and the management layer communicate with each other. The intelligent operation and maintenance system for micro-service application can realize unified application operation and maintenance, can capture abnormal states of applications in real time and actively, can reduce time consumption caused by abnormal analysis, can accurately locate the root cause of faults, and is convenient for recovering system faults. The application can guarantee high availability of the system, save operation and maintenance labor cost, reduce economic losses caused by system faults, and accelerate intelligent transformation.

Description

Technical Field

[0001] This invention belongs to the field of system operation and maintenance technology, and in particular relates to an intelligent operation and maintenance system, method, device and terminal for microservice applications. Background Technology

[0002] Currently, computer and network technologies are developing rapidly, and internet applications are becoming increasingly widespread, covering various industries such as basic services, business transactions, online entertainment, and public services. According to IDC predictions, by 2022, 90% of applications will adopt a microservices architecture, and 35% of product applications will be cloud-native. Service providers need to build enterprise-level cloud-native capabilities in a platform-based manner. In the process of building application system platforms, IT operations and maintenance play a crucial role in improving the overall platform operation and service level.

[0003] IT operations and maintenance (O&M), as a phase in the application system service lifecycle, manages infrastructure and business components through a series of processes, technologies, and methods to provide assurance in terms of quality, cost, efficiency, and security. IT O&M has evolved from manual O&M to automated O&M and DevOps (an automated process practice between a group of software development and O&M teams), but it has never changed the fundamental fact that it is "based on human-defined rules."

[0004] AIOps (Artificial Intelligence for IT Operations) applies artificial intelligence to the IT operations field. Based on existing operations data (application information, system logs, monitoring metrics, etc.), it uses machine learning to deeply address problems that automated operations cannot handle. AIOps does not rely on manually defined rules; it advocates for machine learning models to automatically and continuously learn from massive amounts of operations data, constantly refining and summarizing knowledge about operations scenarios. The machine learning model can be seen as the central hub of AIOps, responsible for directing the collection of necessary operations data, making real-time analyses and decisions, and specifying corresponding scripts and modules for execution, thereby achieving the overall goal of intelligent operations systems.

[0005] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0006] (1) Existing technologies are unable to achieve real-time and efficient online anomaly detection in the face of application system architectures with high complexity and dynamics and massive operation and maintenance data, making it impossible to capture abnormal system states in a timely manner and predict potential faults.

[0007] (2) When dealing with faults in complex systems, existing technologies cannot perform more comprehensive and accurate fault location. Maintenance personnel have difficulty clearly understanding the location and cause of the fault, which hinders subsequent fault recovery and causes further losses. Summary of the Invention

[0008] To address the problems existing in the prior art, this invention provides an intelligent operation and maintenance system, method, device, and terminal for microservice applications.

[0009] This invention is implemented as follows: an intelligent operation and maintenance method for microservice applications, comprising:

[0010] First, the data collector is used to collect the normal execution trajectory and multi-dimensional indicator data of the application system. The detector is used to train the model through the execution trajectory data. The analyzer is used to construct the normal instance correlation graph through multiple indicators. The corresponding execution trajectory and graph are stored in the operation and maintenance database.

[0011] Secondly, real-time execution trajectory data and multi-dimensional indicator data are collected using a data acquisition device, and anomaly detection and interpretation are performed on the current execution trajectory using a detector.

[0012] Then, the analyzer is used to construct a graph related to the current instance, and the current graph is optimized by comparing it with the graphs related to historical instances in the operation and maintenance database.

[0013] Finally, starting with the abnormal results, a random walk is performed in the optimized graph to locate and output the root cause list.

[0014] Furthermore, the intelligent operation and maintenance method for microservice applications includes the following steps:

[0015] Step 1: By integrating Prometheus and Zipkin on the basis of service orchestrator and service mesh, execution trajectory data and multi-dimensional indicator data are acquired, so as to collect multi-dimensional operation and maintenance data for system status analysis.

[0016] Step 2: Obtain normal execution trajectory data of the application system and construct an anomaly detection model; use the normal execution trajectory data of the application system to train the anomaly detection model to obtain a model that can monitor the system's operating status.

[0017] Step 3: Use the trained anomaly detection model to detect real-time execution trajectory data to determine whether there are any anomalies in the application system, so as to achieve the purpose of real-time monitoring of the system's operating status and capturing anomaly information; if an anomaly is found, proceed to step 4.

[0018] Step four: Find the most relevant reference sequences from the database for differential analysis, and output interpretable detection results to make fuller use of relevant anomaly information;

[0019] Step 5: Simultaneously, based on the acquired multidimensional indicator data, locate the root cause of the anomaly and output a root cause list. Store the execution trajectory data, multidimensional indicator data, and generated graph data for system analysis.

[0020] Furthermore, in step one, the acquisition of execution trajectory data and multi-dimensional indicator data by integrating Prometheus and Zipkin on the basis of service orchestrator and service mesh includes:

[0021] Based on the service orchestrator and service mesh Istio, Prometheus and Zipkin are integrated. Zipkin is used to integrate with Istio to collect and store execution trajectory data. Prometheus is used to integrate with Istio to obtain data through push / pull methods, and the data is organized and cleaned according to certain rules.

[0022] Simultaneously, Zipkin's memory storage is used to store the processed execution trajectory data into a new time series; Prometheus's memory storage is used to store multidimensional indicator data.

[0023] Furthermore, in step five, locating the root cause of the anomaly based on the acquired multidimensional indicator data and outputting a root cause list includes:

[0024] (1) When the application system is abnormal, obtain the system properties of the host, container and service instances in the current application system, construct the instance relationship through the correlation of the properties, and generate the instance correlation graph;

[0025] (2) Determine whether the instance correlation graph can be optimized based on historical data. If the instance correlation graph cannot be optimized, proceed to step (4); otherwise, proceed to step (3).

[0026] (3) Aggregate and remove redundancy from the generated instance-related graph, and find similar anomalies from the database to optimize the weights in the current graph so that the walking process can effectively refer to relevant historical anomalies;

[0027] (4) Based on the optimized instance correlation graph and the interpretable detection results, fault location is performed through a random walk strategy, and a root cause list is output.

[0028] Furthermore, in step five, storing the execution trajectory data, multidimensional indicator data, and graph data includes:

[0029] The system utilizes a database to store execution trajectory data; a time-series database to store multidimensional indicator data; and a graph database to store the generated graph data.

[0030] Furthermore, the intelligent operation and maintenance method for microservice applications also includes:

[0031] Instances are created to inject different types of faults in a semi-automatic manner, and test cases are used to verify whether the faults introduced are as expected. If they are, a fault version of the relevant instance is generated, and the fault version is released so that the relevant faults can be reproduced using test cases.

[0032] The intelligent operation and maintenance system for microservice applications provided in this embodiment of the invention includes:

[0033] The infrastructure layer is used to manage hardware resources through containerization technology;

[0034] The application layer is used for developing applications using microservices technology.

[0035] The management layer is used to collect data from the infrastructure and application layers; it is also used for the registration, scheduling, and operation and maintenance management of application instances.

[0036] The management team includes:

[0037] The data acquisition module is used to collect execution trajectory data, resource and application monitoring indicator data during application runtime; it is also used to organize and clean the collected data.

[0038] The anomaly detection module is used to perform real-time anomaly detection based on execution trajectory data using anomaly detection algorithms and anomaly difference interpretation algorithms.

[0039] The fault location module is used to locate the root cause of a fault based on monitoring indicator data.

[0040] The operations and maintenance database module is used to store execution trajectory data, monitoring indicator data, and graph data used in the process of anomaly detection and fault location.

[0041] The infrastructure layer, application layer, and management layer communicate with each other.

[0042] Another object of the present invention is to provide a computer device including a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the intelligent operation and maintenance method for microservice applications.

[0043] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the intelligent operation and maintenance method for microservice applications.

[0044] Another objective of this invention is to provide an information data processing terminal for implementing the intelligent operation and maintenance system for microservice applications.

[0045] Based on the above technical solutions and the technical problems solved, please analyze the advantages and positive effects of the technical solution to be protected by this invention from the following aspects:

[0046] First, addressing the technical problems existing in the prior art and the difficulty in solving them, this paper closely analyzes, in conjunction with the technical solution to be protected by this invention and the results and data obtained during the research and development process, how the technical solution of this invention solves the technical problems, and the inventive technical effects brought about by solving these problems. The specific description is as follows:

[0047] This invention enables unified application operation and maintenance: it monitors operation and maintenance through a three-tier system architecture, collects operation and maintenance-related execution trajectories, resource indicators, application indicators and other data in various forms, conducts comprehensive monitoring and analysis of cross-domain applications, realizes unified and centralized management of multiple applications, and reduces operation and maintenance costs.

[0048] This invention enables real-time anomaly detection: it can actively capture abnormal states of applications in real time, reducing the time consumed by anomaly analysis.

[0049] This invention enables precise fault location: it can accurately pinpoint the root cause of a fault, facilitating system recovery.

[0050] Second, considering the technical solution as a whole or from a product perspective, the technical effects and advantages of the technical solution to be protected by this invention are specifically described as follows:

[0051] The present invention can ensure the high availability of the system, save maintenance manpower costs, reduce economic losses caused by system failures, and accelerate intelligent transformation.

[0052] Third, as supplementary evidence of the inventive step of the claims of this invention, it is also reflected in the following important aspects:

[0053] The technical solution of this invention solves a technical problem that people have long desired to solve but have been unable to achieve:

[0054] Microservices, containerization, and other cloud-native technologies have become the mainstream way to build application systems. The complexity of systems has shifted from the internal layers of services to the service interaction and system architecture layers. Traditional manual operation and maintenance, DevOps and other operation and maintenance methods not only cause high human resource costs, but also make it difficult to predict system failures. When a failure occurs, it is difficult to find the cause of the failure in time, resulting in huge economic losses. There is an urgent need for an intelligent operation and maintenance method.

[0055] This invention focuses on the design and implementation of anomaly detection and fault location in operation and maintenance. It automates system status monitoring, promptly captures anomalies, and performs comprehensive and accurate fault location after occurrence. Maintenance personnel then use the analysis results to initiate subsequent fault recovery processes, thus achieving a more intelligent operation and maintenance approach. Furthermore, it overcomes the current limitations in anomaly detection regarding real-time performance, scalability, and interpretability, and addresses the challenges in fault location regarding accuracy, comprehensiveness, and stability. Attached Figure Description

[0056] Figure 1 This is an architecture diagram of an intelligent operation and maintenance system for microservice applications provided in an embodiment of the present invention;

[0057] Figure 2 This is a sequence diagram of module interactions in an intelligent operation and maintenance system for microservice applications provided in an embodiment of the present invention.

[0058] Figure 3 This is a flowchart of an intelligent operation and maintenance method for microservice applications provided in an embodiment of the present invention;

[0059] Figure 4 This is a schematic diagram of the data acquisition implementation architecture provided in an embodiment of the present invention;

[0060] Figure 5 This is a schematic diagram of the anomaly detection implementation process provided in an embodiment of the present invention;

[0061] Figure 6 This is a schematic diagram of the fault location implementation process provided in the embodiment of the present invention.

[0062] Figure 7 This is an application system data flow and architecture diagram provided in the embodiments of the present invention.

[0063] Figure 8 This is a data processing flowchart corresponding to the application system scenario provided in the embodiments of the present invention.

[0064] Figure 9 This is a diagram illustrating the anomaly detection effect of the application system provided in this embodiment of the invention.

[0065] Figure 10 This is a diagram showing the effect of constructing a fault location map for an application system provided in an embodiment of the present invention.

[0066] Figure 11 This is a diagram illustrating the effect of application system fault location and fault analysis provided in this embodiment of the invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0068] I. Explanatory and Illustrative Embodiments. To enable those skilled in the art to fully understand how the present invention is specifically implemented, this section provides an explanatory and illustrative description of the embodiments described in the claims.

[0069] like Figures 1-2 As shown in the figure, the intelligent operation and maintenance system for microservice applications provided in this embodiment of the invention includes three layers: infrastructure, application, and management. The infrastructure layer manages hardware resources through technologies such as containerization; the application layer develops applications using technologies such as microservices; and the management layer performs operations such as registration, scheduling, and operation and maintenance on application instances. The three layers communicate with each other, and the management layer collects data from the infrastructure layer and the application layer.

[0070] The management layer provided in this embodiment of the invention also includes: an intelligent operation and maintenance center; the intelligent operation and maintenance center collects status data from the infrastructure layer and the application layer, and provides anomaly detection and fault location services for system applications published on the platform to ensure system availability.

[0071] The intelligent operation and maintenance center provided in this embodiment of the invention is mainly divided into four sub-modules: a data acquisition module, an anomaly detection module, a fault location module, and an operation and maintenance database module. The data acquisition module collects runtime data of the application, including execution trajectory data, resource and application monitoring indicator data, etc. The anomaly detection module performs real-time anomaly detection on the system based on the execution trajectory data, including anomaly detection algorithms and anomaly difference interpretation algorithms. The fault location module locates the root cause of the fault based on the monitoring indicator data, including algorithms for constructing relationship graphs between application services and random walk location algorithms. The operation and maintenance database module is responsible for storing the data used in the anomaly detection and fault location processes, including execution trajectory data, monitoring indicator data, and graph data.

[0072] The implementation of the intelligent operation and maintenance system for microservice applications provided in this embodiment of the invention includes two stages:

[0073] During the preparation phase, the data collector gathers normal execution trajectories and multi-dimensional indicator data of the system, the detector trains the model using the execution trajectory data, the analyzer constructs normal instance correlation graphs using multiple indicators, and stores the corresponding execution trajectories and graphs into the operation and maintenance database.

[0074] During the usage phase, the data collector collects real-time execution trajectory data and multi-dimensional indicator data. After the detector performs anomaly detection and interpretation on the current execution trajectory, the analyzer constructs a correlation graph of the current instance. It optimizes the current graph by comparing it with the correlation graph of historical instances in the operation and maintenance database. Finally, starting from the abnormal results, it performs a random walk in the optimized graph to locate and provide a list of root causes.

[0075] like Figure 3 As shown, the intelligent operation and maintenance method for microservice applications provided in this embodiment of the invention includes the following steps:

[0076] S101 integrates Prometheus and Zipkin on the basis of service orchestrator and service mesh. It uses Zipkin to integrate with Istio to collect and store execution trajectory data. It uses Prometheus to integrate with Istio to obtain data through push / pull methods and organizes and cleans the data according to certain rules.

[0077] S102, acquire the normal execution trajectory data of the application system and construct an anomaly detection model; train the anomaly detection model using the normal execution trajectory data of the application system;

[0078] S103, Use the trained anomaly detection model to detect the real-time execution trajectory data and determine whether there is an anomaly in the application system; if there is an anomaly, proceed to step S104.

[0079] S104: Find the most relevant reference sequences from the database for differential analysis and output interpretable detection results; at the same time, locate the root cause of the anomaly based on the acquired multidimensional indicator data and output a root cause list.

[0080] S105: Use a database to store execution trajectory data; use a time series database to store multidimensional indicator data; and use a graph database to store the generated graph data.

[0081] In step S104, the method provided in this embodiment of the invention for locating the root cause of an anomaly based on the acquired multidimensional indicator data and outputting a root cause list includes:

[0082] (1) When the application system is abnormal, obtain the system properties of the host, container and service instances in the current application system, construct the instance relationship through the correlation of the properties, and generate the instance correlation graph;

[0083] (2) Determine whether the instance correlation graph can be optimized based on historical data. If the instance correlation graph cannot be optimized, proceed to step (4); otherwise, proceed to step (3).

[0084] (3) Aggregate and separate the generated instance-related graphs, and find similar anomalies from the database to optimize the weights in the current graphs so that the walking process can effectively refer to relevant historical anomalies;

[0085] (4) Based on the optimized instance correlation graph and the interpretable detection results, fault location is performed through a random walk strategy, and a root cause list is output.

[0086] The intelligent operation and maintenance method for microservice applications provided in this invention embodiment further includes: applying a fault injection method, which, through manipulation of the code and related basic components of the microservice architecture system, introduces different types of faults into different parts of the entire system in a semi-automatic manner, generating a series of related system fault versions. The following information needs to be considered when constructing a fault injection strategy:

[0087] (1) Fault type: The entire system has multiple instances such as basic components and business components. The fault type is defined according to the interaction and function of the instances in the system.

[0088] (2) Fault injection location: Select the appropriate injection location based on different fault types and the actual system environment;

[0089] (3) Fault injection method: The system generates faults during operation by changing the execution code or configuration.

[0090] (4) Expected abnormal behavior, the impact of successful fault injection on the system, i.e., abnormal changes in the execution trajectory and multi-dimensional indicator data of this system.

[0091] Instances are created by modifying code or environment configurations, and faults are introduced into the application system. Test cases are used to verify whether the faults are as expected. If they are, a fault version of the relevant instance is generated. In the subsequent detector and analyzer verification phase, the fault version can be released and test cases can be used to reproduce the relevant faults.

[0092] The specific methods employed by the data acquisition module, anomaly detection module, fault location module, and operation and maintenance database module of the management layer provided in this embodiment of the invention are as follows:

[0093] The data acquisition module is responsible for acquiring execution trajectory data and multi-dimensional indicator data from the system. This is achieved by integrating Prometheus and Zipkin onto a service orchestrator (Kubernetes) and service mesh (Istio). Figure 4Service instances are encapsulated as Pod containers by the service orchestrator. The service mesh assigns an Envoy proxy to each instance for communication. Mixer, as a component of Istio, is responsible for providing policy control and telemetry data. Zipkin integrates with Istio to collect execution trajectory data, provides in-memory storage services, and offers API calls and queries. Prometheus integrates with Istio to acquire data through both push and pull methods, performs data processing and cleaning according to certain rules, and stores the results in a new time series. The collected data has two uses: one is for alarm services, and the other is to provide data manipulation through PromQL.

[0094] The anomaly detection module trains its detection model by learning from normal execution trajectories and provides data storage services. During real-time detection, the trained model examines sequences. When an anomaly is detected, the most relevant reference sequence is retrieved from the database for differential analysis, outputting interpretable detection results. The specific process is as follows: Figure 5 .

[0095] The fault location module provided in this embodiment of the invention locates the root cause of a fault based on monitoring indicator data, including an algorithm for constructing a relationship graph between application services and a random walk location algorithm. When the detector detects a system anomaly, the analyzer obtains the system attributes of the host, container, and service instances in the current system. It constructs instance relationships through attribute correlations, generating an instance correlation graph for fault location analysis. Simultaneously, it optimizes the graph using historical data. Based on the optimized instance correlation graph and the anomaly detection results obtained by the detector, it performs fault location using a random walk strategy. The process is as follows: Figure 6 Graph operations are divided into aggregation and separation operations. The purpose of aggregation is to extract the characteristics of instances, while separation is to remove information irrelevant to the anomaly from the current graph to extract the anomaly characteristics. Then, similar anomalies are searched from the database to optimize the weights of the current graph, so that the walk process can effectively refer to relevant historical anomalies.

[0096] The operation and maintenance database module provided in this embodiment of the invention is responsible for storing data used in the anomaly detection and fault location process, including execution trajectory data, monitoring indicator data, and graph data. The operation and maintenance database module is also responsible for storing data generated during the operation of the detector and analyzer. For execution trajectory data in the detection phase, Zipkin provides in-memory storage and can also persist the data through MySQL; for multi-indicator data, Prometheus also provides in-memory storage and uses InfluxDB for remote storage; and for the generated graph data, it is stored using Dgraph.

[0097] II. Application Examples. To demonstrate the inventiveness and technical value of the technical solution of this invention, this section provides application examples of the technical solution of the claims on specific products or related technologies.

[0098] The intelligent operation and maintenance system for microservice applications provided in this embodiment of the invention is applied to the AIOps of microservice application systems for anomaly detection and fault location; it can meet the needs of platform-based and intelligent operation and maintenance scenarios of microservice applications, and perform real-time and efficient online anomaly detection and more comprehensive and accurate fault root cause location.

[0099] This invention patent has been successfully applied to internally developed smart park application systems, etc. The system is developed using a microservice architecture, such as the data flow architecture diagram of the smart park-road monitoring subsystem. Figure 7 As shown, for Figure 8 In pedestrian and vehicle monitoring and recording scenarios, functional tests were conducted on anomaly detection and fault location. The anomaly detection test results are as follows: Figure 9 As shown, the test results for fault location during the diagrammatic framework phase are as follows: Figure 10 As shown, the test results during the fault analysis phase are as follows: Figure 11 As shown.

[0100] III. Evidence of the Relevant Effects of the Embodiments. The embodiments of the present invention have achieved some positive effects during research and development or use, and indeed possess significant advantages compared to existing technologies. The following description, in conjunction with data, charts, and other materials from the experimental process, illustrates these advantages.

[0101] To demonstrate the effectiveness and advancement of this invention, a comparative analysis was conducted with relevant cutting-edge methods, primarily focusing on anomaly detection and fault localization. Experimental verification was performed using the open-source microservice systems Sock Shop and TrainTicket.

[0102] In the anomaly detection phase, this invention compares the real-time performance, scalability, and interpretability of the MEPFL and TraceAnomaly methods, respectively. The evaluation metrics selected are accuracy, precision, and recall, constructed using a confusion matrix. Experimental results are shown below:

[0103] Table 1. Verification of the real-time performance of anomaly detection in the Sock Shop open-source system.

[0104]

[0105] Compared with the other two methods, the method of this invention reduces the time and cost of anomaly detection while ensuring high accuracy and recall.

[0106] Table 2 Verification of the extensibility of anomaly detection in the Sock Shop and TrainTicket open-source systems

[0107]

[0108] The method of this invention manages two systems in a unified manner. In different systems, the accuracy and false alarm rate of the method of this invention are better than the other two methods, and it has good scalability.

[0109] Table 3. Verification of the interpretability of anomaly detection in the Sock Shop open-source system.

[0110]

[0111] The interpretability of this invention is not as good as MEPFL, because MEPFL uses three supervised methods to classify and extract anomalies in more detail. However, it is an improvement over the unsupervised method TraceAnomaly, especially in relatively complex system environments.

[0112] In the fault location phase, the method of this invention is compared with the CloudRanger and Microscope methods, and the evaluation metrics are set as Mean Accuracy (MAP) and Mean False Alarms (MFP). The experimental results are shown below:

[0113] Table 4 Comparison of fault location methods in Sock Shop scenarios

[0114]

[0115]

[0116] Since interactive failures are usually direct failures, the average accuracy and false alarm rates of the various methods are not significantly different. This invention optimizes the indirect relationships between services by considering metrics, which is an improvement over the other two methods. It also considers the state relationships between service instances, resulting in a significant improvement in handling multi-instance failures.

[0117] Table 5 Comparison of fault location methods in the TrainTicket scenario

[0118]

[0119] In relatively more complex system environments, optimizing graph operations using historical data yields greater gains, with more significant improvements in average accuracy and average false alarms.

[0120] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0121] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An intelligent operation and maintenance method for microservice applications, characterized in that, The intelligent operation and maintenance method for microservice applications includes: First, the data collector is used to collect the normal execution trajectory and multi-dimensional indicator data of the application system. The detector is used to train the model through the execution trajectory data. The analyzer is used to construct the normal instance correlation graph through multiple indicators. The corresponding execution trajectory and graph are stored in the operation and maintenance database. Secondly, real-time execution trajectory data and multi-dimensional indicator data are collected using a data acquisition device, and anomaly detection and interpretation are performed on the current execution trajectory using a detector. Then, the analyzer is used to construct a graph related to the current instance, and the current graph is optimized by comparing it with the graphs related to historical instances in the operation and maintenance database. Finally, starting with the abnormal results, a random walk is performed in the optimized graph to locate and output a list of root causes; The intelligent operation and maintenance method for microservice applications includes the following steps: Step 1: Integrate Prometheus and Zipkin on top of the service orchestrator and service mesh to acquire execution trajectory data and multi-dimensional indicator data; Step 2: Obtain normal execution trajectory data of the application system and construct an anomaly detection model; train the anomaly detection model using the normal execution trajectory data of the application system. Step 3: Use the trained anomaly detection model to detect real-time execution trajectory data to determine if there are any anomalies in the application system; if an anomaly is found, proceed to step 4. Step four: Find the most relevant reference sequences from the database for differential analysis and output interpretable detection results; at the same time, locate the root cause of the anomaly based on the acquired multidimensional index data and output a root cause list. Step 5: Store the execution trajectory data, multi-dimensional indicator data, and generated graph data; In step one, the acquisition of execution trajectory data and multi-dimensional indicator data by integrating Prometheus and Zipkin on the basis of service orchestrator and service mesh includes: Based on the service orchestrator and service mesh Istio, Prometheus and Zipkin are integrated. Zipkin is used to integrate with Istio to collect and store execution trajectory data. Prometheus is used to integrate with Istio to obtain data through push / pull methods, and the data is organized and cleaned according to certain rules. Simultaneously, Zipkin's memory storage is used to store the processed execution trajectory data into a new time series; Prometheus's memory storage is used to store multidimensional indicator data; In step four, the root cause of the anomaly is located based on the acquired multidimensional indicator data, and a root cause list is output, including: (1) When the application system is abnormal, obtain the system attributes of the host, container and service instances in the current application system, construct the instance relationship through the correlation of the attributes, and generate the instance correlation graph; (2) Determine whether the instance correlation diagram can be optimized based on historical data. If the instance correlation diagram cannot be optimized, proceed to step (4); otherwise, proceed to step (3). (3) Aggregate and separate the generated instance-related graphs, and find similar anomalies from the database to optimize the weights in the current graphs so that the walking process can effectively refer to relevant historical anomalies; (4) Based on the optimized instance correlation graph and the interpretable detection results, fault location is performed through a random walk strategy, and a root cause list is output.

2. The intelligent operation and maintenance method for microservice applications as described in claim 1, characterized in that, Step five, storing the execution trajectory data, multidimensional indicator data, and graph data, includes: The system utilizes a database to store execution trajectory data; a time-series database to store multidimensional indicator data; and a low-latency graph database to store the generated graph data.

3. The intelligent operation and maintenance method for microservice applications as described in claim 1, characterized in that, The intelligent operation and maintenance method for microservice applications also includes: Instances are created to introduce different types of faults in a semi-automatic manner, and test cases are used to verify whether the faults introduced are as expected. If they are, a fault version of the relevant instance is generated, and the fault version is released so that the relevant faults can be reproduced using test cases.

4. An intelligent operation and maintenance system for microservice applications, implementing the intelligent operation and maintenance method for microservice applications according to any one of claims 1 to 3, characterized in that, The intelligent operation and maintenance system for microservice applications includes: The infrastructure layer is used to manage hardware resources through containerization technology; The application layer is used for developing applications using microservices technology. The management layer is used to collect data from the infrastructure and application layers; it is also used for the registration, scheduling, and operation and maintenance management of application instances. The management team includes: The data acquisition module is used to collect execution trajectory data, resource and application monitoring indicator data during application runtime; it is also used to organize and clean the collected data. The anomaly detection module is used to perform real-time anomaly detection based on execution trajectory data using anomaly detection algorithms and anomaly difference interpretation algorithms. The fault location module is used to locate the root cause of a fault based on monitoring indicator data. The operations and maintenance database module is used to store execution trajectory data, monitoring indicator data, and graph data used in the process of anomaly detection and fault location. The infrastructure layer, application layer, and management layer communicate with each other.

5. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the intelligent operation and maintenance method for microservice applications as described in any one of claims 1-3.

6. A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the intelligent operation and maintenance method for microservice applications as described in any one of claims 1-3.

7. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the intelligent operation and maintenance system for microservice applications as described in claim 4.

Images