Managing Cloud Costs with Kubecost

Talk Overview

In this talk at DevOps Days Singapore 2025, I shared insights on using Kubecost to optimize cloud costs for Kubernetes workloads, specifically focusing on EKS clusters and GPU-intensive workloads at Invideo.

Key Topics Covered

• Kubecost Implementation: Setting up and configuring Kubecost for EKS clusters
• Cost Visibility: Gaining granular insights into resource consumption and costs
• GPU Workload Optimization: Strategies for efficient GPU utilization and cost management
• Auto-scaling Strategies: Right-sizing resources based on actual usage patterns
• FinOps Best Practices: Implementing cost optimization workflows in DevOps teams
• Real-world Results: Cost savings and efficiency gains achieved at Invideo

Key Takeaways

• How to identify and eliminate wasteful spending in Kubernetes clusters
• Best practices for GPU workload cost optimization
• Actionable strategies to reduce EKS operational costs
• Building a cost-conscious culture in engineering teams

Target Audience

This talk is designed for:

• DevOps Engineers managing Kubernetes infrastructure
• Platform Engineers working with EKS
• FinOps practitioners looking to optimize cloud costs
• Engineering leaders responsible for infrastructure budgets
• Anyone working with GPU-intensive workloads on Kubernetes

About the Event

DevOps Days Singapore is a premier DevOps conference in Southeast Asia, bringing together practitioners, thought leaders, and innovators to share knowledge and best practices in DevOps, Cloud, and modern infrastructure management.