At AWS re:Invent, Amazon introduced EKS Auto Mode, marking a significant expansion in AWS management of Kubernetes clusters, moving beyond the cluster itself. While EKS Auto Mode addresses some long-standing pain points in Kubernetes management, it's important for you to fully understand its capabilities and limitations.
EKS Auto Mode provides some useful high-level features, including streamlining Kubernetes cluster management by providing production-ready clusters without a lot of operational overhead. It also adds or removes nodes in your EKS cluster automatically, based on the demands of your K8s applications, minimizing the need for you to plan capacity manually and making it easier to ensure the availability of your applications. In terms of efficiency, EKS Auto is designed to optimize compute instances (aligning to the NodePool and workload requirements) you have defined, and terminate unused instances. It can improve cost efficiency by consolidating workloads onto other nodes.
EKS Auto mode uses Amazon Machine Images (AMIs) to improve security. These AMIs are treated as immutable for your nodes; they enforce locked-down software, provide read-only root file systems, and enable SELinux mandatory access controls. Nodes launched by EKS Auto mode have a maximum lifetime of 21 days and are automatically replaced with new nodes, ensuring that you regularly cycle your nodes. EKS Auto Mode also keeps your Kubernetes cluster, nodes, and components up to date with the latest patches. It respects any Pod Disruption Budgets (PDBs) and NodePool Disruption Budgets (NDBs) you have configured and includes Kubernetes and AWS cloud features that you’d normally manage as add-ons, including Pod IP address assignments, Pod network policies, local DNS services, GPU plug-ins, health checkers, and EBS CSI storage. You can also customize your NodePools and NodeClasses, although you can’t edit default NodePools and NodeClasses.
EKS Auto Mode also provides managed versions of several critical Kubernetes components, including:
These automated features simplify cluster operations and reduce the operational overhead, which can be especially helpful for teams new to Kubernetes or those with limited resources.
While the automation is welcome, the pricing model may raise some eyebrows. The cost depends on the type and duration of the Amazon EC2 instances that are managed and launched by EKS Auto Mode. In addition to the price of the EC2 instances, the management fee is an additional 12%, which seems high, particularly for teams that see this as core functionality that should be included by default. For organizations running large clusters, this additional cost can quickly add up.
EKS Auto Mode, while useful, doesn't cover all the bases. Fairwinds Managed Kubernetes-as-a-Service already does all of this for our clients, but we also do a lot more to make sure our clients can focus on their core differentiators, not their infrastructure. Our team of site reliability engineers (SREs) architects, builds, and manages all aspects of Kubernetes infrastructure for our clients, while EKS leaves some important K8s management with its customers, including:
One of the most significant limitations of EKS Auto Mode from the Fairwinds perspective is the lack of personalized support:
Fairwinds Managed Kubernetes-as-a-Service offers dedicated support channels, including Slack and ticketing systems, staffed by our own Kubernetes experts. This human touch can be invaluable when troubleshooting complex issues or optimizing cluster performance. In the chart below, we’ve outlined the benefits you can expect from EKS Auto Mode and GKE Autopilot, comparing it to what we offer at Fairwinds with our white-glove Managed K8s-as-a-Service offering:
While EKS Auto Mode provides guardrails for organizations using EKS, it can be restrictive for organizations that need fine-tuned control over their infrastructure. It provides limited ability to override default configurations and your customization options are restricted compared to what Fairwinds provides for our clients. We offer a balance of guardrails and flexibility, allowing for complete customization when necessary while still maintaining best practices for Kubernetes security, cost efficiency, and scalability.
"We appreciate the partnership with Fairwinds for helping us manage our multitude of clusters and providing the guidance to achieve our internal Kubernetes maturity initiatives. Their expertise allowed us to accelerate our Kubernetes journey while ensuring our systems are secure, efficient, and compliant."
- Technology Leader, Food Company
EKS Auto Mode focuses primarily on infrastructure components. However, Fairwinds clients benefit from a more comprehensive approach to cluster management, including:
For organizations operating in multi-cloud environments, EKS Auto Mode is limited to AWS only. Fairwinds Managed Kubernetes-as-a-Service can deliver consistent experiences across AWS, GCP, and Azure, simplifying operations for teams working across multiple cloud platforms.
EKS Auto Mode is an improvement, automating some aspects of Kubernetes; it may be particularly beneficial for small-scale deployments or teams new to Kubernetes that just want to get started. However, for larger enterprises or those with complex requirements, EKS Auto Mode may fall short. These organizations benefit from everything Fairwinds Managed Kubernetes-as-a-Service offers our clients, including:
While EKS Auto Mode is a valuable addition to the Kubernetes ecosystem, it's essential to evaluate your specific needs and consider all available options. For many organizations, a fully managed Kubernetes service provider that goes beyond basic automation and provides expert support and customization is still the optimal choice.
