StormForge Simplifies Cluster Management Using Amazon EKS Auto Mode

StormForge provides automated Kubernetes resource management for customers across industries. The company uses machine learning to analyze Kubernetes cluster data, helping teams rightsize their pod-level resource allocation. Amazon EKS Auto Mode provided a new and innovative way to manage Kubernetes infrastructure, combining secure, scalable cluster infrastructure with integrated AWS-managed Kubernetes capabilities, including automated node-level scaling. Using Amazon EKS Auto Mode, StormForge would no longer need to manage the deployment, scaling, and upgrading of all of its workloads, resulting in decreased time spent on cluster upgrades.

When Amazon EKS Auto Mode was announced, StormForge was interested in exploring the possible benefits of using the service. “We felt there was a very strong synergy between what we do on the application side and what Amazon EKS Auto Mode does on the node side,” says Erwin Daria, director of technical alliances at StormForge. After meeting with AWS solutions architects and seeing a demo, StormForge quickly realized that Amazon EKS Auto Mode would integrate seamlessly with its existing application stack. Excited by the fit, the team decided to migrate its critical workloads.

StormForge followed GitOps best practices in incrementally migrating its self-managed nodes to the Amazon EKS Auto Mode nodes. “To begin with, we migrated approximately 10 percent of our workloads to Amazon EKS Auto Mode,” says Brent Eager, principal software engineer at StormForge. “Once those were stable, we migrated everything else over pretty quickly.” Over a period of 3–4 weeks, the company migrated a total of around 20 critical workloads to Amazon EKS Auto Mode.

StormForge uses Elastic Load Balancing (ELB) to automatically distribute incoming application traffic across multiple targets and virtual appliances in one or more Availability Zones within its Amazon EKS cluster. Because ELB is included as an Amazon EKS Auto Mode feature, StormForge no longer needs to manage ELB separately. “The only potential hang-up was the question of whether a specific ELB feature that we depend on was available out of the box. We reached out to AWS, and they confirmed it was available and explained how to set it up,” says Eager. “In the end, it was like a light switch—we just flipped it on.”

Before implementation, the company primarily managed its own node groups. Using Amazon EKS Auto Mode, StormForge significantly reduced the operational overhead of managing node groups. “Less effort to maintain those clusters is definitely a big win,” says Eager. Now, StormForge engineers spend less time on cluster maintenance and more time on product development. Additionally, because Amazon EKS Auto Mode runs on the control plane rather than cluster nodes, StormForge reduced its cluster pod count by 25 percent. The reduced count alone helped the team optimize resources through a reduction in infrastructure costs and a smaller security perimeter.

Since migrating, StormForge has reduced its cluster upgrade time drastically. “Using Amazon EKS Auto Mode, we’ve reduced the time that we need to do an actual cluster upgrade by 65 percent,” says Eager. That’s because Amazon EKS Auto Mode handles upgrades automatically. “It’s not just a one-time thing at the beginning—it basically manages the upgrades itself,” says John Platt, chief innovation officer at StormForge. “Using Amazon EKS Auto Mode, we can spend more time maintaining customer-facing applications instead of infrastructure.” It also improves availability by removing the need to monitor manual upgrades, such as issues with applications coming back online. With faster, more efficient upgrades, StormForge rarely needs to decommission or recreate clusters. “Not only do we create clusters more efficiently, but they’re also really quick to spin up and get to an operational state,” says Daria. “It’s been a nice change all around in terms of efficiency.”

Since implementing Amazon EKS Auto Mode to optimize its architecture, StormForge has realized an infrastructure cost savings of 30–40 percent. “To us, AWS has always been the obvious place to go to if you’re a company looking to build software,” says Daria. “Technologies like Amazon EKS Auto Mode significantly reduce the complexity of implementing these cloud-focused application solutions.”

The company now educates current and prospective customers on compute optimization. “Amazon EKS Auto Mode is now a principal tentpole in our workshop series,” says Daria. “It’s the foundation for everything.”

As an AWS Partner, StormForge looks forward to continuing to help customers optimize their infrastructure by rightsizing, streamlining, and modernizing their application stacks using Amazon EKS Auto Mode and other AWS services. “Most small, mid-size, and even large enterprises have not made the shift away from traditional monolithic applications on virtual machines,” says Daria. “There are tremendous benefits in using any technology that makes it simpler to implement solutions, and we’re excited to help our customers make that transition.”