Running Kubernetes at LinkedIn

LinkedIn has transitioned much of its workload to Kubernetes, originally relying on a custom container runtime and scheduler created in-house over a decade ago. This shift was prompted by the maturity and scalability of Kubernetes and the increasing complexity of maintaining their legacy systems. Amit and Ronak emphasize that while not all workloads are currently on Kubernetes, the company is actively working to migrate the majority, making efforts to host stateless, stateful, and batch workloads on the platform. They express a desire to achieve an environment where the past challenges faced from their own custom solutions are minimized and smoother transitions are possible.