IaC Cost estimation in your PR

Terracotta AI helps DevOps and platform engineers understand the cost implications of proposed infrastructure changes in a pull request. Terracotta analyzes resource additions, size changes, and drift re-creations to estimate monthly and annual cost impact.

tc:cost can be enabled in the repository UI to run automatically on every PR, or triggered on demand by commenting tc:cost in any pull request. Cost analysis also runs automatically as part of tc:plan.

---

💵 When Cost Analysis Runs

Terracotta supports in-depth, scenario-based cost estimation when asked directly in a PR comment. This includes projections such as:

• 12-month cost forecasts
• Total cost across multiple added resources
• Comparisons between instance types or configurations

Example prompt: @try-terracotta in this diff, what is the cost impact of my changes and what would it look like 12 months out.

Cost analysis runs when:

• tc:cost is triggered manually in a PR comment
• tc:cost is enabled in the repository UI (auto-runs on PR open)
• A plan file is generated during tc:plan (cost is included automatically)
• Resources are added, deleted, or resized
• Drift correction may lead to temporary cost overlap (e.g., re-creating EC2 instances)

---

📊 What the Report Contains

1. Summary Table: Cost Estimates (Dynamic Based on Question)

Terracotta analyzes the plan and calculates deltas for new and modified resources:

---

🧠 AI Comment Content

The AI-generated cost comment includes:

• Breakdown of new vs existing resources
• Dynamic estimates (monthly by default, 12-month if asked)
• Identification of transient overlaps (e.g., during create-before-destroy)
• Warnings about public exposure risks tied to cost (e.g., unused public instances)
• Optional savings recommendations (e.g., consider t3.micro instead of t2.micro)

---

🧮 How Estimates Are Calculated

Terracotta calculates costs based on usage assumptions and region inferred from your Terraform configuration:

• Region: inferred from provider config (e.g., us-west-2)
• Runtime: 730 hours/month
• Pricing model: On-Demand pricing

If pricing data is unavailable:

“Pricing data is not available for this usage type or region.”

---

💡 Cost Optimization Tips

Terracotta may suggest:

• Replacing instance types (e.g. t2.micro → t3.micro, t4g.micro)
• Adding lifecycle policies to reduce drift-driven billing overlap
• Removing unused public resources
• Consolidating underutilized volumes

---

🔇 Zero-Delta Suppression

Terracotta automatically suppresses cost findings when changes don't affect billing-relevant attributes. For example, modifying tags, descriptions, or security group rules on an EC2 instance won't generate a cost finding because those changes don't change the instance's billing profile. This eliminates noisy false positives and keeps the cost report focused on real spending impact.

---

🌐 Pricing Data Source

Cost estimates are powered by an MCP (Model Context Protocol) AWS Pricing server that queries the AWS Pricing API in real-time. This ensures pricing data is always current and region-accurate.

---

📋 TL;DR

• Use tc:cost to trigger cost analysis on demand, or enable it in the UI to run automatically on every PR
• Estimates monthly and annual deltas from proposed changes
• Zero-delta suppression eliminates noise when changes don't affect billing
• Detects overlap risks and offers optimization guidance
• Output is included in the same PR comment as security, drift, and best-practice findings