Skip to content
Calcrivo

Spot Instance Savings Calculator

Price EC2 Spot capacity against on-demand and discount the saving by the work lost to interruptions.

Inputs

USD/hour
%

Spot typically runs 60–90% below on-demand, but the price floats with spare-capacity supply.

instances
hours
events

AWS gives a two-minute termination notice; the Spot placement score and interruption-rate tables estimate this.

minutes

Time to re-run whatever was in flight and re-warm the replacement.

Monthly Saving vs On-Demand

$981.12

Spot Rate per Hour

$0.0288

Spot Fleet Cost

$420.48

Same Fleet at On-Demand

$1,401.60

Instance-Hours Lost to Interruptions

4.0

Discount After Interruption Waste

70.0%

Step by step

  1. Values used

    On-demand rate per hour = 0.0960 USD/hour; Spot discount vs on-demand = 70 %; Instances in the Spot fleet = 20 instances; Hours running per month = 730 hours; Interruptions per instance per month = 2 events; Work lost per interruption = 6 minutes

  2. Spot Instance Savings

    spot rate = on-demand × (1 − discount); effective discount = 1 − (spot cost ÷ useful hours) ÷ on-demand rate, where useful hours exclude work lost to interruptions.

  3. Monthly Saving vs On-Demand

    = 981.12

  4. Spot Rate per Hour

    = 0.0288

  5. Spot Fleet Cost

    = 420.48

  6. Same Fleet at On-Demand

    = 1,401.60

  7. Instance-Hours Lost to Interruptions

    = 4.0

  8. Discount After Interruption Waste

    = 70.0

How it works

Spot bills at a floating market rate for spare capacity, so the headline saving is straightforward — but you also pay for the partial work destroyed when AWS reclaims an instance with two minutes' notice. Dividing the Spot bill by the hours that actually produced finished work gives the effective discount, which is the figure to compare against a Savings Plan. Spot only pays off for interruption-tolerant work: at a few interruptions a month the effective discount barely moves, but on a long non-checkpointed job the rework can erase the entire saving.

Formula

Spot Instance Savings

spot rate = on-demand × (1 − discount); effective discount = 1 − (spot cost ÷ useful hours) ÷ on-demand rate, where useful hours exclude work lost to interruptions.

discount
Headline Spot reduction against on-demand
lost hours
Interruptions × rework time, converted to instance-hours
useful hours
Paid hours that produced completed work

Guides that use this calculator

Frequently Asked Questions

How is Spot Instance Savings calculated?

spot rate = on-demand × (1 − discount); effective discount = 1 − (spot cost ÷ useful hours) ÷ on-demand rate, where useful hours exclude work lost to interruptions. Spot bills at a floating market rate for spare capacity, so the headline saving is straightforward — but you also pay for the partial work destroyed when AWS reclaims an instance with two minutes' notice. Dividing the Spot bill by the hours that actually produced finished work gives the effective discount, which is the figure to compare against a Savings Plan.

Why does Spot Instance Savings matter?

Spot only pays off for interruption-tolerant work: at a few interruptions a month the effective discount barely moves, but on a long non-checkpointed job the rework can erase the entire saving.

What values do I need to enter?

This calculator takes 6 inputs: On-demand rate per hour, Spot discount vs on-demand, Instances in the Spot fleet, Hours running per month, Interruptions per instance per month, Work lost per interruption. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.

What workloads should never run on Spot?

Anything that cannot be killed with two minutes' notice: single-node stateful databases, licence-locked software, and long batch jobs without checkpointing. Diversifying across many instance types and Availability Zones in a Spot fleet is the standard mitigation for everything else.

You might also need