| CPC H04L 67/60 (2022.05) [H04L 41/0631 (2013.01); H04L 41/069 (2013.01); H04L 41/145 (2013.01); H04L 43/50 (2013.01)] | 16 Claims |

|
1. A chaos engineering experimentation method for a system, comprising:
determining target service request traffic, wherein the target service request traffic is minimum experimental traffic of a service request required by a chaos engineering experiment, and the chaos engineering experiment is configured to detect a disaster tolerance capability of the system in a production environment;
acquiring initial service request traffic of a canary cluster; wherein the canary cluster is a cluster formed by servers in a to-be-tested server cluster, the servers being used for canary release when a service is upgraded, and the canary cluster belongs to the production environment;
upon determining that the initial service request traffic of the canary cluster is greater than the target service request traffic, setting part of service request traffic of the canary cluster as service request traffic of a non-canary cluster, so that the service request traffic of the canary cluster is reduced to the target service request traffic; or upon determining that the initial service request traffic of the canary cluster is less than the target service request traffic, setting part of the service request traffic of the non-canary cluster as the service request traffic of the canary cluster, so that the service request traffic of the canary cluster is increased to the target service request traffic; wherein the non-canary cluster is a part other than the canary cluster in the to-be-tested server cluster; and
executing an experimental task of the chaos engineering experiment in the canary cluster, and outputting an experimental result.
|