AWS Fargate Monthly Cost Estimator
Estimate monthly and yearly costs for ECS/EKS workloads running on AWS Fargate.
Note: Rates are sample defaults for estimation and may differ by region/date. Data transfer, load balancers, logging, and other AWS services are not included.
How AWS Fargate pricing works
Fargate pricing is based primarily on the compute resources you reserve and how long those resources run. Instead of managing EC2 instances directly, you pay for vCPU-hours and memory GiB-hours per running task (ECS) or pod (EKS). This model is simple, but costs can grow quickly when tasks run continuously.
At a high level, your monthly Fargate bill is the sum of:
- vCPU usage cost
- Memory usage cost
- Optional extra ephemeral storage cost (above included baseline)
- Any additional AWS service costs (not part of this calculator)
Formula used in this calculator
This page uses a straightforward estimate formula:
- Total monthly hours = hours/day × days/month
- Total vCPU-hours = running tasks × vCPU per task × monthly hours
- Total GiB-hours = running tasks × memory GiB per task × monthly hours
- Total extra storage GiB-hours = running tasks × extra storage GiB × monthly hours
Then each usage bucket is multiplied by its unit price. If you choose Spot, the calculator applies a discount percentage to the compute+storage subtotal.
What this estimator includes
Included
- Linux x86 default pricing profile
- Linux ARM/Graviton default pricing profile
- Windows x86 default pricing profile
- On-Demand and Spot calculation modes
- Monthly and annualized totals
Not included
- Application Load Balancer or Network Load Balancer pricing
- CloudWatch Logs ingestion/retention costs
- NAT Gateway and data transfer charges
- ECR storage and image pull transfer costs
- Regional tax and support plan differences
Example scenario
Suppose you run 10 tasks, each with 1 vCPU and 2 GiB memory, 24 hours/day for 30 days. The calculator converts that into monthly compute usage and multiplies by default rates. If you then switch to Fargate Spot and apply a 50% discount assumption, you can instantly compare your estimated savings.
This is especially useful when deciding between:
- always-on services vs scheduled workloads,
- x86 vs ARM migration,
- On-Demand baseline + Spot burst strategy.
Cost optimization tips for Fargate
1) Right-size task definitions
Over-allocating vCPU and memory is one of the most common budget leaks. Start with usage metrics and tune task sizes incrementally.
2) Use ARM where possible
If your containers support ARM architecture, Graviton-based Fargate options can significantly reduce cost while maintaining strong performance.
3) Combine On-Demand + Spot
Run critical baseline services on On-Demand and place interruption-tolerant jobs on Spot capacity.
4) Schedule non-production environments
Development, QA, and demo stacks often do not need to run 24/7. Automated start/stop schedules can reduce monthly spend immediately.
5) Watch hidden adjacent costs
Even if Fargate compute is optimized, NAT, data transfer, and logging can dominate total container platform spend. Track full-stack cloud cost, not just task runtime.
Final thoughts
A Fargate pricing calculator helps you move from guesswork to informed planning. Use it for early architecture decisions, budget forecasting, and optimization reviews. For production budgeting, always validate assumptions against the official AWS Pricing pages and your own workload telemetry.