What’s the best way to track AWS resources, usage, and costs?

Infrastructure as code is the biggest improvement. Keep Terraform, Pulumi, or CDK configurations in source control so you have a clear record of what should exist, who manages it, and how to remove it when it’s no longer needed.

Tools such as Steampipe or cloud-security platforms can provide a broader inventory and configuration view across services. They’re especially useful when you want to find forgotten resources or configuration problems without manually checking each service.

I use CDK for most configuration and keep projects separated into stacks, and sometimes separate accounts when a project is substantial. That makes the inventory more organized, while the command line is usually faster than clicking through the console for routine checks and troubleshooting.

Use tagging policies and enforce them with organizational controls. Tags such as environment, owner, application, and cost center make it much easier to filter resources and break spending down by project. Infrastructure as code works well alongside this.

Infrastructure-as-code state isn’t always a complete inventory. A resource can sometimes be created successfully even though the provisioning command reports an error, leaving it outside the state file while it still generates charges. A scheduled script that scans every region for instances, databases, and other billable resources provides a useful safety net. Cost Explorer is helpful too, but it may not be real-time.