AWS re:Invent 2024 - Kubernetes
Back to all Re:Invent 2024
Table of Contents
Analytics
 |
The presentation discusses optimizing Amazon EKS for analytics workloads, aligning tools and practices to foster organizational growth, and empowering data engineers by providing APIs to enable autonomy while maintaining platform control. The presenters showcase how AppsFlyer leveraged EKS, Karpenter, and Graviton to reduce costs, improve SLAs, and enhance observability for their Spark-based data processing.
|
Compute
 |
This session presents a comprehensive solution for empowering stateful workloads on Amazon EKS using AWS storage services. The solution demonstrates how to architect a reliable, scalable, and performant real-time inference application that leverages asynchronous processing, Keda, and Karpenter to ensure efficient resource utilization and cost optimization.
|
 |
This session provides a deep dive into Karpenter and introduces Amazon EKS Auto Mode, a new offering that simplifies Kubernetes infrastructure management by offloading cluster operations to AWS. The presentation covers the benefits of Karpenter, the challenges it addresses, and how EKS Auto Mode builds on Karpenter's principles to further streamline Kubernetes workloads.
|
 |
Amazon EKS Auto Mode is a major evolution for easily running production-ready Kubernetes clusters on AWS. It provides fully automated and Kubernetes conformant managed compute, networking, and storage for any EKS cluster, reducing operational overhead and enabling customers to focus on building applications that drive innovation for their business.
|
 |
The video discusses the future of Kubernetes on AWS, highlighting the evolution of Amazon EKS and how it is addressing the challenges faced by enterprises in adopting and managing Kubernetes. The speaker also shares insights from Snowflake's experience in building a robust AI infrastructure on top of EKS.
|
 |
This session provides an overview of how Amazon EKS (Elastic Kubernetes Service) can help customers running a hybrid strategy, with workloads deployed at the edge, on-premises, and on AWS. The session covers the benefits and architectural details of EKS on AWS Outposts, EKS Anywhere, and the newly announced EKS Hybrid Nodes, which allows customers to run Kubernetes nodes on their existing on-premises hardware while offloading control plane management to AWS.
|
 |
This talk focuses on building production-grade resilient architectures with Amazon EKS. It covers the management, observability, and governance aspects of operating a fleet of EKS clusters, emphasizing the use of GitOps, proactive monitoring, and policy-driven governance to ensure consistency and reliability at scale.
|
 |
The video discusses the launch of a new AWS feature called EKS Hybrid Nodes, which allows customers to use their on-premises infrastructure as capacity in EKS clusters. The speaker also covers the challenges of running Kubernetes on-premises, the architecture and networking requirements of EKS Hybrid Nodes, and a customer case study from Northwestern Mutual.
|
Developer Experience
 |
The video discusses how AWS re:Invent 2024 will focus on high-performance generative AI on Amazon EKS. It covers the challenges of running generative AI workloads at scale, the benefits of using EKS, and customer success stories of deploying generative AI on EKS.
|
 |
The session discusses how to build scalable platforms using Amazon EKS, focusing on establishing trust through transparency, collaboration, and empathy. The speakers emphasize the importance of abstracting complexity, leveraging open-source tools, and cultivating a collaborative culture to enable developers to deploy applications quickly and reliably.
|
Networking
 |
This talk provides an overview of networking strategies for Kubernetes, focusing on simplifying cluster operations using Amazon EKS, exposing applications externally with the AWS Load Balancer Controller, and enhancing network resiliency and observability with Istio Ambient Mesh and the Kubernetes Gateway API. The presenters demonstrate these capabilities through a series of hands-on demos, highlighting the benefits and challenges of each approach.
|
Observability
 |
This talk discusses observability strategies for Amazon EKS workloads, covering the importance of logs, metrics, and traces, as well as the various tools and services provided by AWS to collect, analyze, and gain insights from this data to improve the performance and reliability of Kubernetes-based applications.
|
Security
 |
This session covers how to secure Kubernetes workloads running on Amazon EKS, including cluster-level controls, infrastructure-level security, and application-layer best practices. The speakers discuss new features like Cluster Access Management, Pod Identity, and the use of the open-source Cedar project for advanced authorization policies.
|