Traffic forwarding method and apparatus, service system, computing device and storage medium

Abstract

The invention discloses a traffic forwarding method and device, a service system, a computing device and a storage medium used in an experimental example. The method includes: obtaining the attribution information of the online experiment instance from the name service, the attribution information includes the experimental application to which the experimental instance belongs and the cluster to which the experimental application belongs; In the case of an available experimental instance, select an experimental instance from the available experimental instances belonging to the specified experimental application, and forward the traffic request to the selected experimental instance; if there is no available experimental instance belonging to the specified experimental application Next, select an experimental instance from the experimental instances belonging to the cluster to which the specified experimental application belongs, and forward traffic requests to the selected experimental instance. As a result, experiment instances are managed and traffic requests are forwarded according to predetermined experiment rules to ensure service stability to the greatest extent.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a flow forwarding method and device, a service system, computing equipment and a storage medium used in experimental examples. Background technique [0002] In today's era of rapid development of the Internet, massive amounts of information are generated every day, and the way people obtain information from the Internet mainly relies on search and recommendation. Service effect is an important indicator of the success of search and recommendation products, that is, whether it can quickly find out the content that users are most interested in from a large amount of information and push it to users. The improvement of service effect is inseparable from iteration after iteration. Each iteration needs to introduce a small amount of actual traffic to verify the effect. After the effect verification is passed, it can be fully launched. [0003] However, high-frequency eff...

AI Technical Summary

Problems solved by technology

[0003] However, high-frequency effect testing will inevitably bring some problems to system operation and maintenance and service stability: on the one hand, in order to ensure traffic consistency, adding or deleting an experimental version requires the upstream caller to adjust the traffic distribution method

In this way, each effect verification requires upstream cooperation, which increases operation and maintenance costs and reduces work efficiency

[0004] On the other hand, the experimental version itself is to be verified, and there is a high probability of problems with its effect and stability. Once the experimental version fails to provide services normally, the traffic forwarded to the experimental version will be discarded, which will greatly affect the entire System service stability

Images