DevOps is an emerging paradigm to achieve the highly collaboration between system development and operations in order to enable high frequency of software deployment and improve the reliability,stability,elastic and security of production environment.DevOps is typically combined with Cloud computing to realize rapid,on-demand provisioning of underlying resources.Today,an ever-growing amount of DevOps tools,reusable artifacts and Cloud services are available to implement DevOps automation,and a huge number of DevOps knowledge scatters between different communities and sources.As a result,how to make an appropriate decision and select the most suitable method and tools during application design and deployment has become a big challenge.To address this issue,we propose an approach to manage and utilize DevOps knowledge systematically.The approach firstly captures,links,transforms and stores DevOps knowledge from multiple resources in various ways.Then the approach proposes a set of DevOps knowledge taxonomy and implements a knowledgebase prototype.Finally,the approach describes application of DevOps requirements based on predicate logic,and shows how this knowledgebase is utilized.