Savings Plans coverage and utilization: sizing commitments without waste

Savings Plans reward commitment with a discount, but the commitment cuts both ways: you pay for the committed amount whether or not you use it. Two metrics tell you whether a plan is working. Coverage measures how much of your usage gets the discount, and utilization measures how much of your commitment you actually consume. Balancing them is the whole art of buying commitments well.

The two metrics

Utilization is the safety metric. A Savings Plan is a promise to spend a certain amount per hour for 1 or 3 years; if your usage drops below that amount, the unused portion is wasted, you pay for a discount on compute you did not run. Utilization below 100 percent means you over-committed. The first rule of Savings Plans is to keep utilization at or very near 100 percent.

Why you commit to the baseline, not the peak

Coverage is the opportunity metric: usage not covered by a plan pays full on-demand price. It is tempting to raise coverage by committing more, but if you commit above your steady baseline, utilization falls whenever usage dips below the commitment. The resolution is to commit to the floor of your usage, the amount you are confident you will run every hour, so utilization stays near 100 percent, and let the variable capacity above the baseline pay on-demand or run on Spot. This is the layering in the combined Spot and Savings Plans strategy.

The over-commit trap

Suppose your usage averages $10 per hour but ranges from $7 at night to $14 at peak. Committing to $12 per hour looks like good coverage, but every night when usage falls to $7 you waste $5 per hour of commitment, dropping utilization well below 100 percent. Committing to $7 (the floor) instead keeps utilization at 100 percent, and the usage between $7 and $14 pays on-demand or runs on Spot. The floor commitment almost always beats the average or peak commitment on total cost.

Coverage does not need to be 100 percent

A common misconception is that good FinOps means covering all usage with commitments. It does not: a healthy program covers the stable baseline and deliberately leaves the variable top uncovered, because that top is cheaper on Spot or should stay flexible on on-demand. Chasing 100 percent coverage forces over-commitment and tanks utilization. Coverage in the range that matches your baseline, often well under 100 percent for a spiky workload, is the correct target, not a shortfall.

Laddering commitments over time

Because usage grows, commit in layers with staggered end dates rather than one large commitment, so you can adjust as the baseline shifts and avoid a cliff when a big plan expires. As the steady floor rises, add another layer to it. This commitment laddering keeps coverage tracking the growing baseline while protecting utilization, and it hedges against committing too much too early. Compare the plan types in the Savings Plans versus Reserved Instancesguide.

Managing the two metrics

Watch utilization and coverage together: utilization near 100 percent means no wasted commitment, and coverage matched to the baseline means the steady floor gets the discount while the variable top stays flexible. Size new commitments to the demonstrated floor of usage, ladder them as the baseline grows, and resist the urge to cover the peak. Price your steady baseline against the resource catalog so the commitment you buy matches the usage you can prove, keeping both metrics healthy.

FAQ

What to do next

Size commitments to the floor, not the peak. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.