Skip to main content

Usage-based resources

Infracost differentiates the price of a resource from its cost. The price is the per-unit value provided by cloud vendors, while the cost is calculated by multiplying the resource's price by its usage. By default, Infracost displays prices for usage-based resources like AWS S3 or Lambda, providing users with visibility into the few relevant prices (for them) out of the millions of cloud prices.

Furthermore, users have the option to provide usage estimates in a file to calculate costs. This feature also simplifies rapid what-if analysis. For example, they can assess how costs would change if their Lambda function gets 2x more requests, or if they optimize the code to reduce the average run time by 25%.

Name                               Quantity  Unit         Monthly Cost

├─ Requests Cost depends on usage: $0.20 per 1M requests
└─ Duration Cost depends on usage: $0.0000166667 per GB-seconds


Instead of using cloud vendor cost calculators or spreadsheets, you can specify usage estimates in an auto-generated file called infracost-usage.yml. The Infracost GitHub/GitLab App integrations will use this file automatically when it is placed at the repo root, or another location specified in the config file. The CLI can also use this file.

1. Generate usage file

Assuming you have installed the Infracost CLI, use the --sync-usage-file flag to generate a new usage file or update an existing one. You must specify the location of the new or existing usage file using the --usage-file flag:

infracost breakdown --sync-usage-file --usage-file infracost-usage.yml --path /code

This creates/updates the usage file by:

  1. Adding any missing resources or fields as comments with a zero value.
  2. Deleting any resources that are not used in the Terraform project.

When using the --usage-file flag with the breakdown or output commands, cost components with a 0 hourly/monthly quantity are not shown so the output is less noisy. These are included in the JSON format.

2. Edit usage file

Edit the generated usage file with your usage estimates, for example a Lambda function can have the following parameters. This file can be checked into git alongside other code, and updated when needed.

version: 0.1
monthly_requests: 0 # Monthly requests to the Lambda function.
request_duration_ms: 0 # Average duration of each request in milliseconds.

3. Run with usage file

The Infracost GitHub/GitLab App integrations will use the infracost-usage.yml file automatically when it is placed at the repo root. You can also run infracost breakdown or infracost diff with the usage file to see monthly cost estimates:

infracost breakdown --path /code --usage-file infracost-usage.yml

Name Quantity Unit Monthly Cost

├─ Requests 100 1M requests $20.00
└─ Duration 12,500,000 GB-seconds $208.33


Usage profiles

You can use the resource_type_default_usage section of the usage file and create separate files for different traffic profiles, e.g. low/medium/high shown below. This enables you to get a rough estimate for many projects quickly without defining usage values for each individual resource in those projects, by running infracost breakdown --path /code --usage-file infracost-usage-medium.yml.

version: 0.1
monthly_requests: 1000000
request_duration_ms: 100

storage_gb: 5
monthly_write_request_units: 20
monthly_read_request_units: 40

storage_gb: 10
monthly_data_ingested_gb: 10
monthly_data_scanned_gb: 10

Supported parameters

Reference file

The infracost-usage-example.yml reference file contains the list of all of the available parameters and their descriptions for all resource types. These parameters can be added to either a resource (e.g. aws_dynamodb_table.mytable) or a resource type (e.g. aws_dynamodb_table) using the resource type defaults mentioned below.

Resource type defaults

Usage for a resource type, e.g. aws_dynamodb_table, can also be defined in the resource_type_default_usage section of the usage file. Resource type defaults apply to all resources of that type regardless of the module they reside in.

This is useful when you want to create traffic profiles such as lower/medium/high. Resource type defaults can be overridden on a per-resource basis (shown below); usage keys that are re-defined at a resource level override the default, and new usage keys are merged with the defaults.

version: 0.1
storage_gb: 1000 # Set in all DynamoDB table resources

monthly_write_request_units: 200 # Merged with default that defines storage_gb, so both attributes are set for this resource

storage_gb: 50 # Overrides the default

Terraform modules

Usage for resources inside modules can be specified using the full path of the resource. This is the same value as Infracost outputs in the Name column, for example:

version: 0.1
storage_gb: 1000

monthly_requests: 20000
request_duration_ms: 600

Resource arrays/maps

The wildcard character [*] can be used for resource arrays (resources with count meta-argument) and resource maps (resources with for_each meta-argument), such as AWS CloudWatch Log Groups. Infracost will apply the usage values individually to each element of the array/map (they all get the same values). If both an array element such as this[0] (or map element such as this["foo"]) and [*] are specified for a resource, only the array/map element's usage will be applied to that resource. This enables you to define default values using [*] and override specific elements using their index or key.

When wildcard entries exist in the usage file and --sync-usage-file is used:

  • values are generated for each element of the wildcard.
  • entries are added for each wildcard element when usage data is fetched from AWS CloudWatch, which overrides the wildcard value.
version: 0.1
storage_gb: 1000
monthly_data_ingested_gb: 1000
monthly_data_scanned_gb: 200
storage_gb: 1000
monthly_data_ingested_gb: 1000
monthly_data_scanned_gb: 200

EC2 reserved instances

What-if anlaysis can be done on AWS EC2 Reserved Instances (RI) using the usage file. The RI type, term and payment option can be defined as shown below, to quickly get a monthly cost estimate. This works with aws_instance as well as aws_eks_node_group and aws_autoscaling_group as they also create EC2 instances. Let us know how you'd like Infracost to show the upfront costs by creating a GitHub issue.

version: 0.1
operating_system: linux # Override the operating system of the instance, can be: linux, windows, suse, rhel.
reserved_instance_type: standard # Offering class for Reserved Instances. Can be: convertible, standard.
reserved_instance_term: 1_year # Term for Reserved Instances. Can be: 1_year, 3_year.
reserved_instance_payment_option: all_upfront # Payment option for Reserved Instances. Can be: no_upfront, partial_upfront, all_upfront.