Service end-to-end performance monitoring method and device

Abstract

An embodiment of the invention provides a service end-to-end performance monitoring method and device. The method comprises the following steps: obtaining network data of all service components through a test access point (TAP) every a preset first time period, and carrying out filtering and rearrangement on the network data according to pre-configured access control rules to generate original data; according to preset service configuration, carrying out processing on the original data to obtain complete service whole process transaction records and storing the records to a database; and obtaining a current performance index of each service component every a preset second time period according to the complete service whole process transaction records stored in the database. The method anddevice can carry out active, accurate, stable and real-time monitoring on service end-to-end performance to realize fast diagnosis and accurate positioning of faults.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a method and device for monitoring end-to-end performance of services. Background technique [0002] With the rapid development of cloud computing and the Internet, there are more and more businesses, the number of machines has increased by ten times or even dozens of times, and the system architecture has become more and more complex. New technologies have brought investment reduction, performance improvement, resource sharing, and flexibility. It has advantages such as expansion and rapid deployment, but it also brings new challenges and thinking to our operation and maintenance management work. The traditional business monitoring model has been difficult to meet the requirements after cloudification. It is urgent to introduce new operation and maintenance ideas and methods to improve service quality and Operation and maintenance efficiency. [0003] At present, tradition...

AI Technical Summary

Problems solved by technology

[0004] However, due to the increasing number of monitoring machines required after cloudification, the proxy method has serious problems in terms of automation, efficiency, flexibility, stability, real-time and security control. The monitoring data cannot be correlated and needs to be done manually. A large number of log analysis and problem location, the fault location process is complex and inefficient

Although the log method can realize the log association between service components, it is strongly coupled with the business, and the workload of application transformation is heavy, and there is a delay in data output, so it is difficult to achieve full business coverage. come big danger

In addition, in terms of system operation and maintenance, the traditional business monitoring mode is a passive processing method. When a fault occurs, the operation and maintenance personnel generally use the pull-net method to check, collect performance-related logs of each host, and perform aggregation analysis and troubleshooting. This method relies more on personal experience, and it is difficult to locate once encountering tricky problems

Images