When Platform Engineers Lead FinOps: Driving Reliability and $20M in...
summary
Platform engineering drives $20M in 2025 cloud savings while improving reliability across 500+ clusters and 10K+ workloads by combining cost and runtime signals, defining per-workload baselines, and executing targeted changes: separate GPU node pool, move batch Spark jobs to their own pool (average nodes drop ~600→300, monthly cost ~halved), switch Istio to KEDA autoscaling (IstioD pod count −50%), adopt newer instance families, min domains for AZ balance, and fixed 20‑day Carpenter TTL. Establish baselines and views Revisit the legacy Embrace re-architecture
AI-written summary · about
chapters (8)
- 0:00FinOps and platform engineering
- 1:27Agenda and focus
- 2:08Cost reliability disconnect
- 7:04Expedia runtime compute case
- 9:11Baselines and visibility for optimization
- 13:47Legacy systems cost savings
- 16:58Re-architecture for efficiency
- 21:56Key takeaways and closing
related talks
24mPlatform Engineering Is the Key Lever for Sustainable Cloud Infrastructure
12mUnveiling the Future: Navigating Cloud Strategy, Platform Engineering, and Architectural Evolution
44mFinOps at scale: Lessons from scaling Kubernetes workloads
25mPlatform Engineering in Financial Institutions: The Practitioner Panel
30mOptimizing costs across several OpenShift clusters
41mBeyond Uptime: Revolutionizing Fintech Reliability
28mDriving Sustainability with Platform Engineering
06mThe Gold Standard in Platform Engineering: Cloud Foundry :)