Method for automatically adjusting link data sampling rate in real time

Abstract

The invention relates to a link data sampling rate real-time automatic adjustment method, and the method comprises the following steps that a monitoring adjustment end is started, and loads sampling rules corresponding to all services to a cache; the monitoring adjustment end starts a real-time monitoring task to respectively obtain instance names and performance index data of accessed services; the monitoring adjustment end matches the obtained instance name and performance index data of the service with a sampling rule: if matching succeeds, the monitoring adjustment end obtains a corresponding matching sampling rate, and then modifies the current sampling rate of the service into the matching sampling rate, and sends a corresponding modification message to the user and then returns to continue real-time monitoring; if the matching is not successful, the monitoring adjustment end directly continues real-time monitoring. Compared with the prior art, the link data sampling rate can be automatically adjusted in real time on the basis of the sampling rule taking effect in real time, and the service performance cannot be affected.

Description

technical field [0001] The invention relates to the technical field of distributed tracking, in particular to a real-time automatic adjustment method for link data sampling rate. Background technique [0002] With the popularity of the microservice architecture, some problems under the microservice architecture will become more and more prominent. For example, a request will involve multiple services, and the service itself may depend on other services. The entire request path forms a network In the entire call chain, once a certain node is abnormal, the stability of the entire call chain will be affected. In the face of the above situation, at present, some tools that can help understand system behavior and analyze performance problems are usually used , so that when a fault occurs, the problem can be quickly located and solved. These tools are collectively referred to as APM (Application Performance Management, application performance management) tools. Currently, the main...

AI Technical Summary

Problems solved by technology

However, the existing sampling rate for collecting link data is configured in the configuration file of the Agent agent; the above method has the following disadvantages: the sampling rate cannot be modified in real time during the running of the service, and the performance status needs to be manually monitored to determine whether it needs to be modified Sampling Rate

Images