Reserved instance utilization: the discount you have to actually use
A reservation only saves money if it is used. Low utilization or stranded commitments waste the discount you prepaid for. Tracking coverage and utilization, and buying conservatively, keeps commitments paying off. Here is how to manage them.
Quick answer
A reservation or Savings Plan only saves money when it is used. Track two metrics: utilization (are you using the capacity you committed to) and coverage (what share of eligible usage a commitment covers). Low utilization wastes the discount you prepaid; over-commitment strands capacity when workloads change. Buy conservatively to a stable baseline and let on-demand handle the variable top.
Reservations and Savings Plans trade a commitment for a discount, and the discount is only real if the commitment is used. A reservation for capacity you no longer run, or a Savings Plan hourly commitment you do not fill, is prepaid money wasted. Managing commitments is about keeping utilization high and coverage sensible, not just buying discounts.
Two metrics that matter
| Metric | Question | Problem if low/high |
|---|---|---|
| Utilization | Are you using the committed capacity? | Low = wasted commitment |
| Coverage | What share of eligible usage is committed? | Low = paying on-demand you could discount |
High utilization means the capacity you committed to is being used, so the discount is fully realized. Coverage is the complement: of the usage that could be on a commitment, how much is. The goal is high utilization and coverage that matches your stable baseline, without over-committing.
Stranded commitments
The classic waste is a stranded reservation: you committed to an instance type or family, then the workload changed (migrated to a new generation, moved to serverless, shut down), and the reservation no longer matches anything. Prefer the more flexible commitment types (convertible reservations, Savings Plans that apply across instance families) so a workload change does not strand the discount, as covered in reserved instances versus savings plans.
Managing commitments over time
Commit conservatively to the stable baseline you are confident will persist, and let on-demand or Spot handle the variable top, so you never strand a commitment on capacity you stopped using. Track utilization and coverage regularly, ladder commitment terms so they do not all expire at once, and re-evaluate before renewing. Use the break-even discipline from the reserved-versus-on-demand framework to decide how much to commit, then manage what you bought so it stays used.
FAQ
What is reserved instance utilization?
The share of your committed reservation capacity that you actually use. High utilization means the discount you prepaid is fully realized; low utilization means you are paying for committed capacity you are not running, wasting the reservation.
What is the difference between utilization and coverage?
Utilization asks whether you are using the capacity you committed to (low utilization wastes the commitment). Coverage asks what share of your eligible usage sits on a commitment (low coverage means you are paying on-demand rates you could have discounted). You want both high, matched to a stable baseline.
What is a stranded reserved instance?
A reservation that no longer matches any running workload, because the workload migrated to a new instance type, moved to serverless, or shut down after you committed. The prepaid discount is wasted. Flexible commitment types like convertible reservations and Savings Plans reduce this risk.
How do I avoid wasting reservations?
Commit conservatively to the stable baseline you are confident will persist, use flexible commitment types that survive workload changes, track utilization and coverage regularly, ladder terms so they do not all expire together, and re-evaluate before renewing.
Reservations, Savings Plans, or on-demand for the variable part?
Commit (reservations or Savings Plans) to the stable baseline you will use continuously, and let on-demand or Spot handle the variable top. Committing to the variable portion risks stranding the discount when usage drops, so match commitment to the durable baseline only.
Does C3X help manage commitments?
C3X prices your infrastructure from Terraform, so you can size the stable baseline worth committing to before you buy, and estimate the on-demand cost of the variable top. Right-sizing what you commit to is the first step in keeping utilization high.
What to do next
Size the baseline worth committing to before you buy. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.
Share this post
Try C3X on your own Terraform
Free and open source. No API key required. One command to install, one command to estimate.