Monitoring method and device based on feedback type link

Abstract

The invention discloses a monitoring method and device based on a feedback type link. The monitoring method comprises the steps: enabling a server side to extract error description from monitoring information collected by a client side; performing error classification and filtering in sequence according to the error description; and feeding back the filtered error description to the business partyfor analysis and processing. The purpose of the present invention is at least to quickly locate a fault and significantly increase the repair speed of the fault.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a feedback link-based monitoring method and device. Background technique [0002] Now more and more applications are migrating to the cloud-native architecture based on microservices. In the microservice system, with the development of business, the system will become larger and larger, so the call relationship between various services will also be It becomes more and more complicated. A service request will call multiple different microservices to process and return the final result. During this call, the request may fail due to high network delay or transmission error of a certain service. At this time, check the application log. Error analysis would be very time consuming and it would be difficult to quickly locate faults. Therefore, some tools that can help understand system behavior and analyze performance problems are needed, so that when a failure occurs, th...

AI Technical Summary

Problems solved by technology

[0005] There are many errors, and the cause of the problem cannot be quickly identified according to the class method, which leads to low efficiency and cannot quickly locate the problem; because the monitoring system can only obtain the class / method for misclassification, only one class name or method name is used as a sample, resulting in O&M or non-original developers cannot quickly confirm the cause of the error, and may capture a large number of irrelevant exceptions, making it more difficult to identify the cause of the problem, reducing the efficiency of troubleshooting, and failing to quickly locate the problem;

[0006] When categorizing online error descriptions, because of the same error such as dynamic printing problems, and it is impossible to pre-determine which dynamic variable errors are included in the original system, it is impossible to set and automatically classify in advance; the monitoring system is often due to An abnormal dynamic printing problem recognizes similar and similar exceptions as two or more different error categories, and cannot support dynamic configuration of exception analysis and classification templates, which leads to the existence of a large number of similar errors when viewed by the business side , which affects the investigation efficiency of the business side; similarly, it is impossible to accurately alert specific exceptions;

[0007] Existing distributed link monitoring cannot distinguish between business anomalies and system anomalies, and a large number of errors accumulate together to form interference factors; existing distributed link monitoring cannot distinguish whether an error is a business-defined error or a system error. When troubleshooting is required, a large number of business errors and system exceptions are mixed together, which greatly affects the speed of troubleshooting on the business side;

[0008] Online errors cannot be associated with the problem feedback from the business side, only through document retrieval and personal memory; because online errors cannot be associated with the problem feedback from the business side, the same error is repeatedly captured, when the business side needs to troubleshoot the problem , each time you have to rely on memory, rely on locally classified error documents, or analyze from scratch, which is not friendly enough for the business side to troubleshoot problems

It is also not conducive to the business side to quickly locate and solve problems

Images