Accelerating Kubernetes cluster development: With K.I.N.D and Helmfile
I made a resolution to share CFP on Medium this year
Alex Jones, Engineering Director, American Express
Format
Solo Presentation
Speaker
Alex Jones, Engineering Director, American Express
Working for thirteen years within the industry and dedicating the last four to hybrid cloud migrations and open ecosystem adoption, Alex has spoken at meetups for both Start-ups and large corporations that are navigating the cloud native landscape.
Abstract
A relationship with cloud and on-premises infrastructure is becoming a key part of what it means to be a developer. The bar to entry is lowering and enabling deployments, monitoring and operations; but the local development experience is becoming less representative of cloud infrastructure.
When engineers do manage to create local Kubernetes cluster environments, they often fall short of the sophisticated interplay of systems found in their cloud counterparts. Increasing adoption of service mesh and in-cluster governance mechanisms has made the rubric for debugging more complex. With this being compounded by operator style deployments, interoperable components and mixed topology infrastructure.
In this talk we work through the daunting landscape of infrastructure and tooling, identifying areas that would be invaluable to emulate within the local development experience and those that are extremely challenging. Guided by anecdotal experience and examples there will be a presentation of sig/KIND tooling as a stepping-stone to emulate complex cluster topology. Partnered with the Helmfile project to enable complex multi-chart deployments both locally and on production systems.
Presentation Outline (STC)
- Introduction
2. Ambitions of the speaker’s team to become cloud native at American Express.
3. Infrastructure challenges faced within a large organisation with hybrid infrastructure.
4. Tooling challenges around scaling configuration.
5. K.I.N.D introduction and architecture.
6. Helmfile introduction and architecture.
7. How these tools helped us to develop a solution.
8. Demonstration.
9. Q&A.
Benefits to the ecosystem
This talk reinforces that the end-user community is a driving factor behind the focus of the SIG’s that are dedicated to making cloud native the ecosystem more accessible. We narrate a real world scenario where our production Kubernetes clusters have been built with complex and often unknown idiosyncrasies. Working with community projects, emulating these environments has been made trivial and helped to upskill developers in SMI and CNI in tandem.
Key audience learnings
- Acknowledgement that Kubernetes is now rarely just vanilla. There are a multitude of interoperability mechanisms in play that mean a simple local cluster does not necessarily represent a real environment that your organisation operates within.
- Overview and understanding of the benefits of K.I.N.D and the use case for local development.
- Overview and understanding of Helmfile and it’s orchestration capabilities.
- Sample code.
Resources
