Application service autonomous detection and control method, storage medium, equipment and system

Abstract

The invention discloses an application service autonomous detection and control method, a storage medium, equipment and a system, and relates to the technical field of application monitoring, and themethod comprises the following steps: distributing a self-defined inspection interface corresponding to each application to each application; when the application is started, detecting the registration condition of the application by a custom check interface; when the custom check interface detects that the application registration is successful, feeding back a success signal, otherwise, feeding back a failure signal; and terminating the application corresponding to the failure signal according to the failure signal. According to the application service autonomous detection and control method,the application with registration failure can be terminated in time, so that the occupied load is liberated.

Description

technical field [0001] The application monitoring technical field of the present invention specifically relates to an application service autonomous inspection and control method, storage medium, equipment and system. Background technique [0002] Java applications in the field of software engineering often have a large number of service distributed deployment scenarios, that is, multiple instances of the same service will be deployed to achieve load balancing, especially in the case of a large number of service visits, distributed deployment is the design of the system architecture The key is to enable the entire system to load a large number of visits; [0003] In the distributed system architecture, it is particularly important to gracefully start, stop, and upgrade applications. That is, when an application in the cluster is stopped or restarted, other instances of the application need to provide users with continuous and smooth services. There may be problems, because ...

AI Technical Summary

Problems solved by technology

[0003] In the distributed system architecture, it is particularly important to gracefully start, stop, and upgrade applications. That is, when an application in the cluster is stopped or restarted, other instances of the application need to provide users with continuous and smooth services. There may be problems, because when the application is about to stop, it is very likely that a client request will be sent to this instance. If the application is forcibly stopped at this time, this part of the request will be lost, causing a bad experience for the user and even affecting the system. normal function use

[0004] The traditional solution is to have a registration center service that continuously polls each application instance, that is, sends a heartbeat. If an instance does not respond in time, the registration center service will drop the application instance from the gateway. So that the client request is no longer forwarded to the instance. After the instance restarts successfully, the application instance is restored in the gateway after receiving a successful restart signal. However, there is a problem in the traditional solution: relying on the registration center to continuously poll For each application instance, the registration center has a heavy load, and when the registration center finds that the service is unavailable, some requests have already been forwarded to the application instance;

Images