Cost optimization is a massive topic right now, and for good reason. But while the market focuses on cloud costs, there’s a big opportunity for your on-prem workloads too, specifically bare metal.

In this session, we’ll show you how to easily squeeze every last MHz out of your bare metal servers when running Kubernetes. You’ve paid for beast-mode servers, but they might be near idle, locked into running control plane nodes. It’s time to unlock the cage and put that horsepower to work. Whether it’s in data centers or larger edge sites, there’s an exciting opportunity to optimize them that we want to show you.

Using Canonical MAAS and LXD, KVM and Cluster API, we’ll show you how to pool and slice your bare metal server resources to rightsize control plane nodes and workers, with full machine lifecycle management.

By the end, you’ll understand exactly how to unleash the power of the bear (metal):

without paying the Broadcom hypervisor tax

without resorting to the pain of managing heterogeneous compute hardware per site

without incurring the risk of contention from scheduling workloads on control plane nodes