AWS bill line item
Fargate-vCPU-Hours:perCPU
The vCPU-hours and the companion memory GB-hours requested by your container tasks, billed per second from task start to stop. It's metered by resource request, not by underlying node.
Fargate vCPU and memory hours
Verified against official AWS documentation last checked
Why it shows up
Two independent meters (vCPU and memory) mean a task can be cheap on one dimension and expensive on the other; requesting more memory "just in case" inflates the memory line even when actual CPU use stays low.
Spikes when task count scales out, task definitions over-request vCPU/memory relative to real use, or tasks run longer than expected. A stuck shutdown hook keeps billable seconds ticking.
How to cut it
- Right-size task definitions' CPU/memory requests against actual utilization instead of default templates.
- Use a discounted interruption-tolerant capacity option for batch/async workloads.
- Tighten auto-scaling scale-in thresholds so idle capacity doesn't linger.
- Check for tasks stuck in a long-running drain/shutdown state that keep accruing billed seconds.
How to read a usage type
The colon in Fargate-vCPU-Hours:perCPU doesn't introduce an instance type the way it does in a code like BoxUsage:m5.large. This code carries no region or direction token either, so there's nothing to break down here.
Source
Source: AWS docs →Found Fargate-vCPU-Hours:perCPU on your own bill?
Drop your Cost Explorer CSV into the AWS Bill Analyzer. It'll flag line items like this one on your own bill and link straight back to the fix.
Analyze your own bill →A bill read by hand goes stale tomorrow.
Reading one export line by line takes real work. CostMon reads the next one for you, every day, for every provider you connect.