Kasasa Migrates to Amazon EKS for Automatic Cluster Scalability and Upgrades

Overview

Kasasa, a leading financial technology and marketing services company, specializes in helping local banks and credit unions compete with megabanks by providing branded products, marketing solutions, and operational support. To continue their growth and maintain operational efficiency, Kasasa sought to modernize their cloud infrastructure, focusing on improving scalability, automation, and operational resilience. The company was using a self-managed Kubernetes (K8s) environment for deploying and managing applications, but faced challenges related to cluster scalability, upgrades, and operational overhead.

CloudGeometry partnered with Kasasa and AWS to execute a strategic migration of over 200 services from a self-managed Kubernetes cluster to Amazon Elastic Kubernetes Service (EKS). This migration was designed to enhance Kasasa's infrastructure by automating cluster management, scaling, and upgrades while minimizing disruption to ongoing operations.

Opportunity

Kasasa's legacy Kubernetes environment, while functional, lacked the scalability and automation required to support the company's growing needs. The primary challenges included:

• Assessing the current infrastructure and identifying areas for optimization.
• Migrating existing workloads to a more modern, robust, fully managed Kubernetes environment.
• Minimizing disruption to ongoing operations during the migration process.
• Ensuring 100% of changes met or exceeded existing rigorous security and constraints
• Reducing operational overhead associated with Kubernetes management and upgrades.

Kasasa required a solution that would streamline its Kubernetes operations, provide automatic scaling, and ensure seamless updates, all while maintaining their commitment to high availability and performance.

Solution

CloudGeometry adopted a structured, two-phase approach to execute the Kubernetes migration to Amazon EKS. This approach included thorough discovery and planning, followed by a meticulous execution phase to ensure a smooth transition and minimize operational disruption.

Phase 1: Discovery

• AWS Well-Architected Framework Review (WAFR): CloudGeometry began by conducting a comprehensive review of Kasasa's existing infrastructure using AWS's Well-Architected Framework. This review helped identify key optimization opportunities and set the groundwork for the migration plan.
• AWS Well-Architected Tool: to capture and assess workloads, providing detailed insights and recommendations.
• AWS Trusted Advisor: to identify performance improvements and cost optimization solutions.
• Stakeholder Interviews: CloudGeometry interviewed key stakeholders from Kasasa's engineering and operations teams to gather insights into the migration's existing pain points and objectives.
• Automated Infrastructure Discovery: Using automated discovery tools, CloudGeometry assessed Kasasa's current environment, uncovering potential challenges such as outdated configurations and compatibility issues.
• Drift Detection Analysis: A drift detection analysis was performed to identify discrepancies between the current infrastructure and the desired state, ensuring alignment for the migration.
• Migration Planning: Based on the findings from the discovery phase, CloudGeometry developed a comprehensive Kubernetes migration plan that addressed the technical and operational challenges identified.

Phase 2: Migration Delivery

• Infrastructure as Code (IaC): CloudGeometry prepared the necessary infrastructure for Amazon EKS using CloudGeometry’s Open Source CGDevX reference implementation for EKS, enabling a consistent and repeatable deployment process.
• AWS Kubernetes Migration Factory (KMF): CloudGeometry utilized AWS's Kubernetes Migration Factory for pilot migration, testing the new environment with a select group of applications.
• Updating Kubernetes Manifests: As part of the migration, Kubernetes manifests were updated to ensure compatibility with the latest version of Kubernetes (v1.30), addressing changes in APIs and configurations.
• Platform Supporting Services: Datadog, external-dns, nginx ingress controller, metrics server, gitlab runners, and rabbitmq were also upgraded.
• Continuous Delivery (CD) Pipelines: Existing pipelines were reconfigured to allow seamless integration with the new EKS cluster to avoid configuration overhead associated with pipeline reconfiguration.
• Secrets Management Upgrade: Prior to the upgrade, security and compliance tests required multiple iterations as part of the test and staging promotion to production. Introducing integrated granular secrets management and validation upstream in the development process reduced late-stage test failures and rework, and cut time to production release. 
• Cluster Management via GitOps and CD: CloudGeometry implemented cluster management processes through continuous delivery, allowing Kasasa to leverage automated scaling, upgrades, and configuration management in EKS.

Outcome

The migration to Amazon EKS successfully transformed Kasasa's cloud infrastructure, delivering the following benefits:

• Automatic Scalability: Kasasa now benefits from automatic cluster scaling, eliminating the need for manual intervention and enabling the infrastructure to adjust dynamically based on demand, improving cost efficiency and resource utilization.
• Streamlined Operations: With EKS managing the underlying Kubernetes infrastructure, Kasasa has significantly reduced the operational overhead of managing and upgrading clusters.
• Improved Agility: The migration has enabled Kasasa to deploy applications more rapidly and with greater consistency, improving development velocity and operational efficiency.
• Minimized Disruption: Through careful planning, the migration was completed with minimal disruption to Kasasa's ongoing operations, allowing the company to maintain service availability throughout the transition.
• Future-Proof Infrastructure: Kasasa is now equipped with a modern, scalable, and highly available Kubernetes environment that can easily support future growth.

As Sandeep Seshadri, Executive Vice President of Engineering at Kasasa, noted, "By taking advantage of Amazon EKS and CloudGeometry's skilled developers, Kasasa completed the entire migration of over 200 services from a self-managed Kubernetes cluster to AWS in less than two months."

AWS Services Used

• Amazon Elastic Kubernetes Service (EKS): Managed Kubernetes service for simplifying cluster management and providing automatic scaling and upgrades.
• AWS Well-Architected Framework (WAFR): Framework for assessing and optimizing the architecture of cloud workloads
• AWS Kubernetes Migration Factory (KMF): A set of tools and best practices for migrating Kubernetes workloads to AWS.

Other Services Used or Upgraded

• KubeCost: Successfully upgraded the version, ensuring continuity in cost management and optimisation during the transition.
• Datadog: Managed and configured on EKS, maintaining seamless application performance monitoring and log tracking across environments.
• Sumo Logic: version upgrade and integration enhancement enabling improved log analytics and real-time operational insights.
• Terraform: Configured Run Deck on EKS and upgraded Terraform to enhance automation.

Lessons Learned

• Comprehensive Discovery is Essential: A thorough initial assessment is critical to understanding the existing infrastructure and identifying potential migration challenges. Stakeholder interviews and automated discovery tools helped uncover critical insights and challenges.
• Flexible Migration Approach: The pilot migration and incremental approach were key to minimizing risk and ensuring a smooth transition. CloudGeometry's flexibility in adjusting the scope based on discovery findings allowed the team to tackle issues early and refine the approach as necessary.
• Collaborative Project Management: Effective communication between Kasasa, CloudGeometry, and AWS teams was vital for the success of the migration. Daily standups, milestone reviews, and transparent scope modifications kept the project on track.
• Technical Considerations: Updating Kubernetes manifests for compatibility with new versions (v1.30) was essential for ensuring a smooth migration.

Conclusion

Kasasa's migration to Amazon EKS was driven by the need to enhance scalability and cost efficiency. The migration delivered significant operational improvements and cost savings, with 20-30% reductions in infrastructure expenses through optimized resource utilization and automation.

Leveraging CloudGeometry's deep expertise in Kubernetes and AWS cloud adoption, Kasasa successfully transitioned from a self-managed cluster to Amazon EKS, significantly lowering operational overhead and infrastructure maintenance costs. This shift empowers Kasasa's engineering team to focus on innovation rather than infrastructure management, while AWS managed services platform ensures continuous upgrades, patches, and seamless scaling.

More than just a technical migration, this transformation strategically aligns Kasasa's cloud operations with industry best practices for long-term scalability and efficiency, reinforcing its ability to deliver innovative financial technology solutions to its customers.

‍

About CloudGeometry: CloudGeometry delivers expert technical services, helping our clients unlock the full potential of cloud-native open source tooling and commercial platform technologies. As an AWS Advanced Consulting partner, our certified solution architects and platform engineers help address the range of challenges facing enterprise innovators and venture-funded startups alike. The Cloud Native Computing Foundation has accredited us as a Kubernetes Certified Service Provider.

‍

‍