AWS cost mastery: Track, tag, and slash expenses without losing scale

In the fast-paced world of platform engineering, unchecked AWS costs can undermine even the most innovative setups. In this session, Kishan Baranwal explains how his team evolved cost management from ad hoc fixes to a proactive, integrated governance model, yielding substantial savings through rigorous tracking, enforced tagging, and optimized resource allocation. Drawing from large-scale deployments, Baranwal presents a framework that blends technical tactics with organizational changes to embed cost consciousness into platform operations.

The talk covers high-ROI techniques across AWS services, including application profiling for CPU and memory right-sizing, mandatory tags for organizational attribution, and automated alerts for new resource launches. It also highlights migrations such as Intel-to-Graviton for compute efficiency, S3 lifecycle policies for data tiering and deletion, EBS gp3 upgrades, and unused volume cleanup.

Key takeaways:
• Enforce tagging and inventory tracking for complete visibility and accountability across resources
• Use profiling and trends to right-size compute, cutting 20-40% on EC2/ECS without performance hits
• Automate S3/EBS policies to eliminate unused storage and optimize for infrequent access
• Harness spot instances, Graviton, and autoscaling to reduce idle expenses while ensuring uptime
• Build policies and a questioning culture to optimize every decision for sustainable cloud economics