Module 4.4: Cloud FinOps & Infrastructure Economics
Cloud cost anatomy, reservation strategy, spot instance economics, FinOps culture, and team-level cloud budgets.
Lesson 1: Cloud Cost Anatomy
Cloud spending is the fastest-growing line item in most engineering budgets. Understanding the anatomy of cloud costs — compute, storage, network, and managed services — is the first step to optimization.
EC2/GCE/Azure VMs, containers, Lambda functions. The largest cost component. Over-provisioning is rampant because engineers fear performance issues.
Block storage, object storage, database storage. Data grows but is rarely cleaned up. 30-50% of stored data hasnt been accessed in 12+ months.
Data transfer between regions, zones, and to the internet. Often the most surprising cost component because its hard to predict.
Pull your last 3 months of cloud bills. Break down by: compute, storage, network, managed services. Which category has the highest waste?
Lesson 2: Reservation & Commitment Strategy
On-demand pricing is 3-4x more expensive than committed pricing. A proper reservation strategy is the easiest 30-60% cost reduction available.
1-year commitment: 30-40% savings. 3-year commitment: 50-60% savings. Risk: if your needs change, you pay for unused capacity.
AWS Savings Plans / GCP CUDs offer flexibility across instance types. Better for dynamic workloads than traditional RIs.
60-90% cheaper than on-demand. Can be terminated with 2 minutes notice. Perfect for batch jobs, CI/CD, and fault-tolerant workloads.
Analyze your cloud usage: what percentage is consistent (reservable) vs. variable (on-demand)? Calculate savings from committing 60% of your baseline to 1-year reservations.
Lesson 3: FinOps Team & Culture
Cloud cost optimization is not a one-time project — its an ongoing practice. FinOps creates a culture where engineering teams are accountable for their cloud spend.
Tag every resource with team, environment, and project. Allocate 100% of cloud costs to business units. Untagged resources go to a "shame" budget that forces cleanup.
Give each team a cloud budget. They own it. Overspend: justify or optimize. Underspend: reallocate. This creates natural incentives for efficiency.
Automated alerts when spending exceeds baseline by 20%+. Catches: forgotten resources, misconfigured auto-scaling, runaway batch jobs.
Implement cost tagging for your top 5 most expensive cloud services. Set up a daily cost anomaly alert (>20% above 7-day average).