cargo-bill

cargo-bill is a Cargo plugin that estimates AWS Lambda cost and cold-start impact from your compiled Rust binary. It is designed for platform and FinOps workflows where you want cost visibility before deployment.

What it does

• Builds your project in release mode and analyzes binary footprint.
• Estimates Lambda storage and execution cost.
• Includes both Lambda cost components:
  • Compute cost (GB-seconds)
  • Request cost ($0.20 / 1M requests)
• Supports optional AWS Free Tier deductions:
  • 400,000 GB-seconds
  • 1,000,000 requests
• Fetches dynamic pricing via AWS Pricing API with fallback to static rates.
• Predicts cold-start latency from binary size and memory configuration.
• Outputs either table format or JSON for automation.

Installation

From crates.io

cargo install cargo-bill

From install script

curl -sL https://raw.githubusercontent.com/0xBoji/cargo-bill/master/install.sh | bash

Quick start

As a Cargo plugin:

cargo bill lambda

Equivalent binary invocation:

cargo-bill bill lambda

Run with explicit parameters:

cargo bill lambda \
  --region ap-southeast-2 \
  --architecture arm64 \
  --memory 512 \
  --executions 1000000 \
  --include-free-tier

JSON output mode:

cargo bill lambda --json

CLI options

--region <REGION>                     AWS region (default: us-east-1)
--memory <MEMORY_MB>                  Lambda memory in MB (default: 128)
--executions <COUNT>                  Number of invocations (default: 1,000,000)
--architecture <x86_64|arm64>         Lambda architecture (default: x86_64)
--include-free-tier                   Apply monthly free tier deductions
--provisioned-concurrency             Assume warm start (cold start = 0)
--json                                Print machine-readable JSON

Cost model

Estimated execution cost is:

total_cost = compute_cost + request_cost

Where:

• compute_cost = billable_gb_seconds * region_arch_price
• request_cost = (billable_requests / 1_000_000) * 0.20

If --include-free-tier is enabled:

• billable_gb_seconds = max(total_gb_seconds - 400000, 0)
• billable_requests = max(executions - 1000000, 0)

Accuracy notes

• Dynamic pricing uses AWS Pricing API location names mapped from AWS region codes.
• If pricing API access fails, cargo-bill falls back to static Lambda rates.
• On non-Linux hosts (macOS/Windows), compiled binary size may differ slightly from Amazon Linux ELF builds. For stricter production parity, use cargo-lambda or cross for cross-compilation.

Example output

AWS Lambda Cost Estimation Report:
+-----------------------------------------------+
| Metric                              Value      |
+===============================================+
| Binary Size (MB)                    15.52      |
| Architecture                        arm64      |
| Estimated Monthly Storage Cost      $0.0015    |
| Estimated Cost per 1000000 Requests $2.9433    |
| Predicted Cold Start Latency        1862.18 ms |
| Dynamic API Pricing Used            Yes        |
+-----------------------------------------------+

CI and releases

• CI: lint + tests on push and pull request.
• Automated release PRs: release-plz workflow.
• Publishing: crates.io + GitHub release artifacts.

License

MIT