Cloud Cost That Doesn’t Hurt: A Practical FinOps Checklist

Engineering speed and cloud cost don’t have to be enemies. Use this lightweight checklist to trim waste and keep velocity.

Foundations

• Tag everything (owner, env, cost-center)
• Rightsize instances and databases quarterly
• Turn off dev/staging outside work hours

Kubernetes

• Set requests/limits; use VPA for right-sizing
• Autoscale with HPA; consolidate low-traffic workloads
• Prefer spot/preemptible for stateless jobs

Storage & Data

• Lifecycle S3/Blob to infrequent access
• Compress and partition data; drop zombie snapshots
• Monitor data transfer—egress adds up fast

Observability

• Budget alerts and anomaly detection
• Sample logs; store only what you read
• Dashboards for $/service and $/team

Governance

• Use guardrails and policies-as-code
• Review top 10 costly resources monthly
• Make cost part of PR review for infra changes

FinOps works best when it’s cultural: visible, routine, and owned by teams—not a once-a-year firefight.