Introduction

AWS Fargate is a serverless compute engine that simplifies running containerized applications by eliminating the need to manage servers. In this blog you will find a concise breakdown of how AWS Fargate pricing works, making it easy to understand for anyone exploring containerized applications on AWS.

How is AWS Fargate Pricing Calculated?

AWS Fargate is a serverless compute engine that eliminates the need to manage servers, letting you focus on deploying and scaling containerized applications. With Fargate, there are no upfront costs—you pay only for the resources your applications use. This includes compute (vCPU), memory, storage, operating system, and CPU architecture choices. Whether you're running on Amazon Elastic Container Service (ECS) or Amazon Elastic Kubernetes Service (EKS), Fargate’s pricing model provides flexibility and cost efficiency, especially for dynamic or unpredictable workloads.

Here’s a breakdown of the pricing factors:

• vCPU: The virtual CPU allocation required for your containerized application.
• Memory: The memory allocation required for your workloads.
• Operating Systems: Pricing may vary depending on your OS choice.
  • Windows OS is supported on Amazon ECS.
• CPU Architecture: Options such as x86 or ARM can influence costs.
  • ARM architecture is currently supported only for Amazon ECS.
• Storage: You get 20 GB of ephemeral storage for each task or pod by default; additional configured storage incurs extra charges.

AWS Fargate offers a pay-as-you-go pricing model, providing flexibility and cost efficiency for containerized workloads. With no upfront costs or long-term commitments, Fargate allows you to pay only for the vCPU, memory, and storage resources your applications consume. Additionally, Fargate supports both On-Demand pricing for predictable availability and Spot pricing for accessing unused capacity at significantly reduced costs.

Note: AWS Fargate does not offer a free tier.

Pricing details (On-Demand)

All pricing examples in this blog refer to rates for the US East (N. Virginia) region

Linux/X86 pricing:

• per vCPU per hour $0.04048
• per GB per hour $0.004445

Linux/ARM

• per vCPU per hour $0.03238
• per GB per hour $0.00356

Windows/X86

• per vCPU per hour $0.046552
• OS license fee - per vCPU per hour $0.046
• per GB per hour $0.00511175

Fargate Spot Pricing

AWS Fargate Spot allows you to run interrupt-tolerant workloads on spare AWS capacity at a significantly reduced cost—up to 70% lower than regular Fargate pricing. Spot prices are dynamic and adjust based on supply and demand trends.

• Linux/X86 Architecture:
  • vCPU per hour: $0.0124419
  • GB per hour: $0.00136621
• Linux/ARM Architecture:
  • vCPU per hour: $0.00995229
  • GB per hour: $0.0010942

Note: Fargate Spot is currently available only for Amazon ECS workloads using the Linux Operating System on x86 or ARM CPU architectures.

Fargate Ephemeral Storage Pricing

AWS Fargate provides 20 GB of ephemeral storage by default for every task (ECS) or pod (EKS) at no additional cost. If your workload requires more storage, you can configure additional ephemeral storage, which incurs charges based on usage.

Pricing for additional ephemeral storage:

• $0.000111 per GB per hour

Supported CPU and Memory Configurations

Fargate offers flexibility with various CPU and memory combinations to fit your application needs. Here are the supported configurations:

• 0.25 vCPU
  • Memory: 0.5 GB, 1 GB, and 2 GB
• 0.5 vCPU
  • Memory: Min. 1 GB to Max. 4 GB (in 1 GB increments)
• 1 vCPU
  • Memory: Min. 2 GB to Max. 8 GB (in 1 GB increments)
• 2 vCPU
  • Memory: Min. 4 GB to Max. 16 GB (in 1 GB increments)
• 4 vCPU
  • Memory: Min. 8 GB to Max. 30 GB (in 1 GB increments)
• 8 vCPU
  • Memory: Min. 16 GB to Max. 60 GB (in 4 GB increments)
• 16 vCPU
  • Memory: Min. 32 GB to Max. 120 GB (in 8 GB increments)

When Does Billing Start and Stop?

AWS Fargate calculates pricing based on the duration your container runs, measured per second for precision:

• Billing starts as soon as the container image begins downloading (Docker pull).
• Billing ends when the containerized task (Amazon ECS) or pod (Amazon EKS) terminates.

For Linux containers, usage is rounded up to the nearest second with a 1-minute minimum.

For Windows containers, billing follows the same per-second model but has a 5-minute minimum.

Additional Charges to Consider

While AWS Fargate pricing covers vCPU, memory, and storage, you may incur additional charges when using other AWS services or transferring data. Here are some key considerations:

• Amazon CloudWatch Logs: If your containers send logs to CloudWatch for application monitoring and logging, you will be billed for CloudWatch usage.
• Public IPv4 Addresses: Tasks (ECS) or Pods (EKS) using public IPv4 addresses incur additional charges. You can find more details on public IPv4 pricing in the VPC Pricing Page.
• Data Transfer: Any data transferred to or from your containers is billed at standard AWS data transfer rates.

For a full breakdown of pricing for commonly used AWS services, refer to the respective pricing sections on their official AWS service detail pages.

Pricing Example: Calculating Fargate Costs for Windows ECS Tasks

Let’s calculate the monthly cost for 10 ECS Tasks running on Windows for 1 hour (3600 seconds) daily over 30 days. Each task uses 1 vCPU and 2 GB of memory.

1. Monthly CPU Charges
   • Formula: # of Tasks × # of vCPUs × CPU price per second × Duration (seconds per day) × # of days
   • Calculation: 10 × 1 × $0.0000254167 × 3600 × 30 = $27.45
2. Monthly Windows OS Charges
   • Formula: # of Tasks × # of vCPUs × OS price per second × Duration (seconds per day) × # of days
   • Calculation: 10 × 1 × $0.0000127778 × 3600 × 30 = $13.80
3. Monthly Memory Charges
   • Formula: # of Tasks × Memory (GB) × Memory price per second × Duration (seconds per day) × # of days
   • Calculation: 10 × 2 × $0.0000027778 × 3600 × 30 = $6.00

Total Monthly Compute Charges

To get the final cost, add the charges:
$27.45 (CPU) + $13.80 (Windows OS) + $6.00 (Memory) = $47.25

Conclusion

AWS Fargate provides a flexible and cost-efficient solution for running containerized workloads without the need to manage servers. Its pay-as-you-go and Spot pricing models allow you to optimize costs based on workload requirements, whether you need predictable availability or are leveraging spare capacity for significant savings.

AWS Fargate’s transparent pricing structure, combined with its scalability and serverless approach, makes it an ideal choice for modern containerized applications.