codecentric migrated an aging microservice environment at a publicly listed wholesale company in collaboration with Google's Professional Service Organization (PSO). The company now benefits from serverless approaches such as scalability, cost efficiency, and simplified delivery.
The project at a glance
- No disruption to the development process during migration.
- A blueprint was created that solved various problems and thus helps other teams to benefit more quickly from Google Cloud Run.
- Operating costs are expected to be reduced by 65%.
- Happy customer! Three microservices were migrated to Google Cloud Run and put into productive operation instead of the planned one microservice.
Initial situation
The development teams at a large, publicly listed corporation have access to runtime environments via an in-house cloud platform. However, this in-house cloud platform severely limits development teams, especially when it comes to choosing technologies and offering only Kubernetes as a runtime environment. In this case, the Kubernetes cluster resources are shared by the teams. Particularly problematic is the delayed insight into costs, which are only published once a month. As a result, it is difficult to organize workloads cost-effectively.
Google Cloud's technology portfolio offers a wide range of options for runtime environments, real-time cost transparency, and a wide range of other cloud computing services. codecentric worked with a development team on the customer side to test the feasibility of the concept in a pilot project. Particular attention was paid to the migration from Kubernetes to Google Cloud Run as the new runtime environment.
Solution
The migrated product consists of a Springboot microservice environment comprising 20 services. Converting all services to a 100% serverless technology like Google Cloud Functions would have gone beyond the scope of a migration. With Google Cloud Run, however, it was possible to leave the microservice environment, which was already based on containers, on Spring Boot, but still benefit from serverless functionality (scalability, cost efficiency and greatly simplified delivery).
The migration was carried out while the application was being developed as usual. The project started with an enablement and architecture review phase to identify the most important building blocks and implications of Google Cloud Run as a collaboration between the developers on the customer side, Google PSO and codecentric.
What is Google Cloud Run?
Cloud Run is a managed computing platform that enables companies to run containers directly on Google's scalable infrastructure. Google Cloud Run is designed to work with other services in Google Cloud, allowing companies to create full-featured applications.
Source: Google (2022)
Result
In the course of the migration, a total of three microservices were migrated to Google Cloud Run in the pilot project and are already in productive use. A roadmap for the further migration path of all microservices was created and is being implemented. The blueprint created during the migration aids implementation. It supports other product teams in their migration with templates for various problems. In addition, teams can use many code snippets and infrastructure-as-exampled of code. They benefit from these artefacts and can implement further migrations many times more efficiently.
Using Google Cloud Run is expected to reduce the cost of running the development platform by 65%.
Jan Rümenapf
IT Consultant, codecentric AG
Many of our customers run this type of cloud environments or similar ones. The project demonstrated how easy it is to modernize aging microservice projects in order to take advantage of serverless concepts.
Jan Rümenapf
IT Consultant, codecentric AG
Any questions about the project?
Are you interested in a custom cloud-based solution for your company? Let's talk without obligation.
Further projects of codecentric AG
Find out about other successful projects that we have completed with our customers. Perhaps you will find inspiration for a use case in your company here.
Jan Rümenapf