Welcome to the cloud-native era, where managing Kubernetes workloads has been simplified, thanks to Amazon Elastic Kubernetes Service (EKS). With AWS EKS, you can easily run containerized applications, but what if you could push the boundaries even further? Enter AWS Fargate, a serverless compute engine that’s changing the way we approach Kubernetes workloads. In this post, we will take a deep dive into AWS EKS Fargate and how it can alleviate the pain points of Kubernetes management.

Exploring AWS EKS Fargate

AWS EKS Fargate is an integration of the AWS EKS and Fargate services. It allows you to run Kubernetes Pods without having to manage the underlying EC2 instances. With EKS Fargate, each Pod gets its own isolated compute environment, making it a truly serverless solution.

Fargate manages Pod CPU and memory differently from typical Kubernetes operations. When a Pod is scheduled on Fargate, the vCPU and memory reservations within the Pod specification determine how much CPU and memory to provision for the Pod. The requested CPU and memory must equal the limit for all containers in an EKS Fargate Pod, which is a departure from standard Kubernetes, where requests can be lower than the limits.

The Perks of Using EKS Fargate

By eliminating the need to provision and manage EC2 instances, EKS Fargate offers significant advantages. You no longer need to choose server types, decide when to scale the node groups, or optimize cluster packing. This leads to operational efficiency, better resource utilization, and overall cost savings.

A Deep Dive into Fargate Pod Configuration

Scheduling and provisioning Pods in Fargate have a few unique aspects. Firstly, the larger of the vCPU and memory requests out of any Init containers and long-running containers is chosen for the Pod. An extra 256 MB is added to each Pod’s memory reservation for Kubernetes components like kubelet, kube-proxy, and containerd.

Fargate rounds up to a compute configuration that closely matches the sum of vCPU and memory requests, ensuring Pods always have the necessary resources. It offers a range of vCPU and memory combinations, which gives you the flexibility to choose a combination that best suits your workload.

Automated OS Patching in Fargate

One of the significant pain points that EKS Fargate addresses is the handling of OS patching. AWS periodically patches the OS of Fargate nodes to ensure security, eliminating the need for manual intervention. If a Pod is unable to be evicted during the patching process, Amazon EKS sends an event detailing the Pod eviction failure, giving you a chance to handle the issue before the Pod is deleted.

Real-World Scenarios: Moving Clients to EKS Fargate

Through my experience, I’ve seen the transformation that occurs when clients transition from traditional EC2-node based EKS to EKS Fargate. One client was struggling with managing EC2 nodes, especially during traffic spikes. After migrating to EKS Fargate, they were no longer burdened with this operational overhead and could focus more on their application development.

Step-by-Step Guide: How to Use EKS Fargate

If you’re ready to give EKS Fargate a go, here’s a simple guide to get you started:

1. Create a Fargate profile: This determines which Pods should run on Fargate. You specify this by using labels and namespaces.

2. Run a Pod in Fargate: Create a Kubernetes Pod that matches the selectors in your Fargate profile. Remember to specify the Pod resources properly in the Pod manifest file to avoid under or overprovisioning resources.

3. Monitoring and management: You can view your Pods’ logs via Amazon CloudWatch and manage them like in any other Kubernetes environment.

Conclusion

In the rapidly evolving world of cloud computing, EKS Fargate is a game-changer for running Kubernetes workloads. Its ability to eliminate the need for EC2 management and deliver a truly serverless Kubernetes experience is revolutionary. By incorporating EKS Fargate into your Kubernetes operations, you can significantly reduce your management overhead and focus more on what matters most - delivering value to your users. If you’re using EKS and have yet to leverage Fargate, it might be time to consider it.