The Checklist
The Cloud & AI Cost Optimization Checklist
The actionable, do-it-now companion to the Playbook: every lever in one trackable list, grouped by where the waste actually hides.
Last updated 26 items across 6 categories
Compute & rightsizing
Storage & data transfer
AI / LLM spend
Commitments & discounts
Visibility & alerting
Governance & hygiene
Want the reasoning behind these levers? Read the Cloud & AI Cost Optimization Playbook, the narrative guide this checklist is the do-it-now companion to.
Working through the Commitments & discounts group? Run the numbers first with the commitment discount calculator — it shows the break-even utilization and stranded spend before you revisit a reserved instance or savings plan.
Rightsizing a Kubernetes cluster instead of a fleet of plain instances? The Kubernetes cost allocation & idle capacity calculator splits the bill into used, workload idle, and cluster idle before you touch a single request.
Working through "Use prompt caching for repeated context"? How to cut your LLM bill without downgrading your product models caching alongside four other levers as an honest waterfall, including the exact hit rate below which caching costs you money instead of saving it.
Checking off "Tag spend by service, team, or project"? The Cost Allocation & Tag Coverage Scorer turns that into a number — how much of the bill you can actually attribute today, and what it would take to make it chargeback-ready.
Working through "Move cold data to a cheaper storage tier" or "Clear orphaned volumes, snapshots, and IPs"? The AWS storage cost calculator prices S3 as Standard, Intelligent-Tiering, and an explicit lifecycle side by side, then breaks EBS waste into gp2-to-gp3, orphaned-volume, and snapshot-archive savings.
Running Datadog or CloudWatch alongside the rest of this list? The observability cost calculator prices Datadog's host high-water mark, custom-metric cardinality, and log ingest vs. index separately, plus every CloudWatch logs and metrics meter, then ranks the config changes that bring each one down.
Running Snowflake or BigQuery? The data warehouse cost calculator prices Snowflake's idle-before-suspend waste and resume-minimum trap against BigQuery's on-demand-vs-Editions crossover, then ranks the config changes that bring each one down.
Connect a provider. See one number.
Connect your first provider and CostMon normalizes it alongside everything else you run. Your team gets one number everyone can check.