AWS Fargate Cost Estimator
Estimate your monthly Fargate bill based on average usage. Rates are prefilled by region/profile and can be treated as planning estimates.
vCPU-hour: $0.00000 | GB-hour memory: $0.00000 | GB-hour storage (above 20GB): $0.000000 | Public IPv4-hour: $0.00000
Note: This estimate excludes data transfer, load balancers, NAT Gateway, CloudWatch logs/metrics, ECR image transfer/storage, and Savings Plans/Spot discounts.
How this Fargate price calculator works
This calculator estimates monthly AWS Fargate compute costs by combining four billable components: vCPU, memory, additional ephemeral storage, and optional public IPv4 addresses. You enter your average running footprint (tasks, CPU, memory, runtime), and the calculator multiplies that usage by the selected regional rate card.
Inputs used in the estimate
- Average Concurrent Tasks: How many tasks are running at the same time on average.
- vCPU per Task: CPU allocation in each task definition.
- Memory per Task: Memory allocation in GB.
- Ephemeral Storage: First 20GB per task included; only extra storage is charged.
- Public IPv4 per Task: Optional hourly public IPv4 billing if assigned.
- Hours/Day and Days/Month: Runtime schedule for your monthly total.
Fargate pricing formula
The monthly estimate is calculated with this structure:
Total = CPU cost + Memory cost + Extra storage cost + Public IPv4 cost
- CPU cost = Task-hours × vCPU per task × vCPU-hour rate
- Memory cost = Task-hours × Memory GB per task × Memory GB-hour rate
- Extra storage cost = Task-hours × Max(Storage GB − 20, 0) × Storage GB-hour rate
- Public IPv4 cost = Task-hours × IPv4 per task × IPv4-hour rate
Example scenario
Suppose you run 8 tasks continuously, each with 0.5 vCPU and 1GB RAM, and no extra storage or public IPv4 addresses. In this setup, your bill is mostly vCPU + memory time. If you then increase memory to 2GB per task, your memory line item roughly doubles, even if CPU remains unchanged.
That is why right-sizing task definitions is one of the fastest ways to lower container cost on Fargate.
What this estimate includes and excludes
Included
- On-demand Fargate vCPU usage
- On-demand Fargate memory usage
- Ephemeral storage above included baseline
- Public IPv4 hourly charge (if used)
Not included
- Application Load Balancer / Network Load Balancer charges
- Data transfer and NAT Gateway costs
- CloudWatch Logs ingestion and retention
- ECR image storage and transfer
- Fargate Spot pricing and Savings Plans discounts
Practical ways to reduce Fargate spend
1) Right-size CPU and memory
Many teams over-allocate memory "just in case." Use task metrics (CPU/memory utilization and throttling) to tune to realistic limits.
2) Move suitable services to ARM
If your containers are multi-arch compatible, Linux/ARM (Graviton) can reduce compute cost versus x86 in many regions.
3) Scale down outside business hours
For non-24/7 workloads, cutting runtime from 24h/day to 10h/day can immediately cut monthly costs by more than half.
4) Avoid unnecessary public IPv4 assignments
Use private networking with NAT/load balancer patterns where appropriate; this helps control direct IPv4 hourly charges.
Final note
This page is designed as a fast planning tool for engineering and finance conversations. Before committing budget, verify current rates and billing dimensions in the AWS Pricing page and your Cost Explorer data.