FinOps tips for optimizing your cloud infrastructure and data costs
summary
Capital One frames FinOps around a concrete claim: nearly 40% of cloud bill comes from shared data platforms, so cost control must target data not just EC2 or S3. The team builds and ships Slingshot with a query advisor (proprietary SQL optimizer), scheduling recommendations that auto‑size Snowflake warehouses, and managed workflows that apply fixes for users. A single warehouse once spiked 200% week-over-week; pilot deployments cut waste and the tools are rolling company‑wide.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 2:25Agenda
- 2:50About Capital One
- 3:39Our Tech Journey
- 4:36How many spend a lot of money on the cloud
- 5:20FinOps definition
- 7:00Four pillars of FinOps
- 8:51FinOps domains
- 11:42Data
- 11:59Cloud costs are variable
- 13:14User personas
- 15:16Finance personas
- 18:08Shared accountability
- 20:18Proactive notifications
- 22:04Relevant notifications
- 23:17Optimizing workloads
- 24:24Fixing problems
- 27:28Application stack
- 30:44FinOps journey
- 33:50Key takeaways
- 34:29Summary
related talks
25mBuilding a FinOps practice to realise the true value of cloud
45mIron FinOPS
24mAn Introduction to FinOps
29mFinOps: Principles and cloud cost observability
31mGetting the Most out of Every Dollar Spent in Cloud
27mHow to “do FinOps” with CloudHealth by VMware
45mHow to Practically Optimize Cloud Spend with FinOps
48mCloud FinOps & Kubernetes Optimisation at Scale