terraform cost preflight

The problem

Infrastructure as Code allows for rapid deployment, but it also makes it easy to accidentally provision expensive, idle, or over-sized resources. Standard Terraform plans show what will change, but they don't flag the architectural anti-patterns that lead to massive cloud bills.

What it does

• Scans Terraform plans and diffs for nine specific high-cost anti-patterns including NAT gateway sprawl and orphaned EBS volumes.
• Identifies over-provisioned RDS instances, oversized Lambda memory, and missing S3 lifecycle policies.
• Flags unused Elastic IPs, idle load balancers, and CloudWatch log groups lacking retention policies.
• Maps every finding to a specific resource address and defines the cost-vs-availability trade-off.
• Redirects authoritative pricing calculations to Infracost to ensure you never rely on stale, static data.

Frameworks & tools

Terraform, OpenTofu, AWS, GCP, Azure, Infracost.

Why this beats prompting it yourself

A generic prompt often yields vague advice or inaccurate pricing hallucinations. This skill uses a strict semantic layer to catch specific, hard-to-spot patterns while maintaining a clear boundary between architectural logic and real-time pricing data.

Use cases

• Reviewing a Terraform plan for NAT gateway sprawl across multiple availability zones.
• Audit of EBS volumes and snapshots to find unattached or orphaned storage.
• Checking new RDS or Lambda resource blocks for over-provisioning before deployment.
• Ensuring all new CloudWatch log groups include a defined retention period.

Known limitations

Does not calculate authoritative dollar figures. Limited to nine specific anti-patterns; it is not an exhaustive FinOps suite or a real-time monitoring tool.