Skip to content
AWSAmazon EC2Compute

aws_spot_instance_request cost estimation

An EC2 instance run at Spot prices, up to ~90% below on-demand, in exchange for possible interruption. c3x prices it at the on-demand ceiling (a t3.micro is ~$7.59/month).

An aws_spot_instance_request launches an EC2 instance using spare capacity at the Spot price, which floats with supply and demand and runs up to ~90% below the on-demand rate. The catch is interruption: AWS can reclaim the instance with a two-minute warning when it needs the capacity back or when the Spot price exceeds your max.

Because the Spot price floats, c3x prices a spot request at the on-demand rate as a conservative ceiling, a t3.micro is ~$0.0104/hour, ~$7.59/month on-demand, and your actual Spot cost will be meaningfully lower. Pricing the ceiling means estimates never understate cost based on a discount that can change.

Spot is ideal for fault-tolerant, stateless, or checkpointable work, batch jobs, CI runners, rendering, big-data processing, anything that survives a node disappearing. For durable, single-instance workloads, use on-demand or a Reserved Instance instead. And for managed fleets, an Auto Scaling group or EKS/ECS with a spot capacity type handles interruption and replacement better than standalone spot requests.

c3x prices the request from instance_type at the on-demand ceiling, flagging the spot discount as upside.

Terraform example

A minimal but realistic configuration that C3X can estimate.

resource "aws_spot_instance_request" "worker" {
  ami                    = data.aws_ami.al2023.id
  instance_type          = "t3.micro"
  spot_price             = "0.01"
  spot_type              = "one-time"
  wait_for_fulfillment   = true

  tags = {
    Name = "batch-worker"
  }
}

Pricing dimensions

What you actually pay for when you provision aws_spot_instance_request.

DimensionUnitWhat's being charged
Instance (on-demand ceiling)per hourPriced at the on-demand rate as a conservative ceiling; actual Spot price floats up to ~90% lower.
$0.0104/hour on-demand for t3.micro → ~$7.59/month (real Spot cost is lower)
EBS volumeper GB-monthThe instance's root/data EBS volume bills separately.

Sample C3X output

A t3.micro spot request, priced at the on-demand ceiling (real Spot is lower):

aws_spot_instance_request.worker
└─ Instance (t3.micro, on-demand ceiling)   730 hours   $7.59
                                            Monthly     $7.59

Optimization tips

Common ways to reduce aws_spot_instance_request cost without changing the workload.

Use Spot for fault-tolerant, stateless work

Up to ~90% vs on-demand

Spot's up-to-90% discount is real money for batch jobs, CI runners, rendering, and stateless workers that tolerate the two-minute interruption notice. Don't run stateful singletons or latency-critical services on raw spot requests.

Prefer managed fleets over standalone spot requests

Resilience plus spot savings

An Auto Scaling group with a mixed/spot allocation, or EKS/ECS with a spot capacity type, handles interruption, diversification across instance types, and automatic replacement, more robust than individual aws_spot_instance_request resources.

Diversify instance types

Fewer interruptions, lower price

Interruption risk and price vary by instance type and AZ. Requesting from a pool of compatible types (via a fleet) reduces interruptions and catches the cheapest available capacity.

Design for interruption

Avoids rework cost on interruption

Checkpoint long jobs, make workers idempotent, and handle the two-minute Spot interruption notice to drain gracefully, so an interruption costs minutes, not lost work.

FAQ

How much does an EC2 Spot instance cost?

The Spot price floats with supply/demand, typically up to ~90% below on-demand. Because it changes, c3x prices a spot request at the on-demand rate as a conservative ceiling (e.g. ~$7.59/month for a t3.micro), your real Spot cost will be lower.

Why does c3x show the on-demand price for a spot request?

Spot prices float, so any fixed spot number would be wrong by next week. Pricing the on-demand ceiling means the estimate never understates cost based on a discount that can vanish, treat the spot savings as upside.

Should I use a spot request or an Auto Scaling group?

For anything beyond a one-off, prefer a managed fleet, an Auto Scaling group with spot allocation, or EKS/ECS with a spot capacity type, which handles interruption, instance-type diversification, and replacement far better than standalone spot requests.

How does c3x estimate the cost?

From instance_type at the on-demand rate (the conservative ceiling), plus the EBS volume. It flags spot usage so you know the real cost is lower.

Related resources

AWS
aws_instance

On-demand EC2, the ceiling spot is discounted from

Estimate this resource in your own Terraform

Free, open source, no API key. C3X parses your Terraform and shows line-item cost for every resource, including aws_spot_instance_request.

Get StartedStar on GitHub
← Browse all AWS resources