Enhanced Spot Ocean Controller: The fastest and most efficient scaling for your K8s applications

Reading Time: 3 minutes

Fast scaling is a pressing business need in many B2C industries like media distribution, streaming, and gaming. Spot by NetApp is proud to offer the fastest scaling on earth with its enhanced Ocean controller.

Just released: Enhanced Spot Ocean controller for Kubernetes-on-AWS customers

Today, we are excited to release an enhanced Ocean controller for our Kubernetes-on-AWS customers. The enhanced Ocean controller enables even faster scaling, enhanced high-availability features, and a minimized footprint within your clusters.

Background: Why do controllers matter?

Spot Ocean — the world’s best compute engine for Kubernetes on public clouds — relies on a controller to be deployed into your existing K8s cluster. The Ocean controller is a K8s deployment collecting metrics and events about your cluster. This data is then pushed back to a SaaS layer to generate automated actions and AI recommendations.

Once an event such as a pending pod is detected and the information is sent back to our SaaS layer, Ocean immediately triggers scaling events to ensure your clusters always have the infrastructure needed to make the pods’ requirements.

Benefit #1: 60%+ faster pod scaling

The ability to immediately scale up by large numbers is crucial in many online B2C industries, like media streaming and gaming.

To support this need, the enhanced Ocean controller is designed with an event-driven architecture. This allows an even faster execution of scaling events, such as a pending pod in your cluster. With the enhanced controller, our customers have seen over 60% faster scaling times compared to traditional polling-based controllers. This result was consistent in the many use cases and scenarios tested.

Event-driven architecture allows the Kube-apiserver to notify the Ocean controller immediately when events take place. This improves the response time to scaling events and reduces the load placed on the K8s API server, compared to frequently polling the server for recent changes.

Benefit #2: Faster self-healing

We know that availability is your first-and-foremost priority. The last thing you want is for a controller to break just when you need it to kick in.

Therefore, the enhanced Ocean controller is also designed with leader election and multiple replicas out-of-the-box. Having multiple replicas allows one of the pods to serve as the leader in communicating back to the Spot SaaS layer, while the other replicas keep in sync with each other. In the event something was to happen with the leader pod, one of the other replicas would quickly be able to jump in as a new leader.

Benefit #3: Reduced footprint & resource consumption

Controllers consume memory on your kube-apiserver. We know it, and we know it’s a burden.

As a result, the enhanced Ocean controller brings a new gospel with an exceptionally small image size, as low as under 50mb. Smaller image sizes allow for faster deployments. We are also seeing a decrease in memory consumption from the Ocean controller by up to 70% in a cluster with over 30,000 pods.

Benefit #4: Proactive remote support

As with any Kubernetes resource, debugging and troubleshooting is a hassle. We’re here to help with that.

The enhanced Ocean controller 2.0 provides an opt-out capability that allows the controller to continuously publish logs from the controller back to our SaaS layer. Thus, Ocean is able to troubleshoot even faster without any intervention on the user’s end. In fact, we might even be knocking on your door if we can predict a problem before one arises.

Get started with the enhanced Ocean Controller

Are you ready to enjoy the 10X functionality of the new Ocean controller? Check out our docs to make sure you are running the latest version of the Ocean controller.

Are you struggling to manage your K8s on public cloud at scale? Schedule a 1:1 meeting with our razor-sharp Solution Architects today.