Build Your Own Cluster API Provider the Easy Way

The presentation discusses how to build your own Cluster API Provider and highlights common patterns, development and debugging workflows, and common pitfalls to take into account when writing your own provider.

• Cluster API Providers make the experience of provisioning clusters consistent
• Providers handle the infrastructure or environment-specific operations
• Cluster API has higher order functionality like automatic scaling and upgrades
• Building a Cluster API Provider is similar to building a Kubernetes operator
• Tilt is a useful tool for testing and debugging locally

The presenter mentions that if you have built a Kubernetes operator in the past, you should have no problem building a Cluster API Provider. They also discuss the importance of spreading machines across failure domains to prevent the entire cluster from going down if one rack fails.

Over the past year, the adoption of Cluster API (CAPI) has been growing with more end-users using it to provision their clusters. And increasingly it’s being adopted inside commercial products and other OSS projects (e.g. EKS-Anywhere). With this growth comes an increase in the variety of the Cluster API Providers you can choose from. What if none of the existing providers suit your use case? Perhaps you want to bring your own hosts or integrate with a custom infrastructure provisioning mechanism. If that's the case, this talk is for you. Come learn from the maintainers of existing CAPI providers on how to get started creating your own provider. It’s hard but at the same time easier than it sounds. Although every provider has unique considerations in its offering, there is a lot of commonality when it comes to writing a provider. This talk will highlight the common patterns, develop and debug workflows, and common pitfalls / gotchas to take into account when writing your own provider.Click here to view captioning/translation in the MeetingPlay platform!