Cloud cost tools often show you what happened after it’s too late: spend arrives 24–48 hours later, ownership is fuzzy, and the answers live in a separate UI engineer rarely open. In this session, we’ll share how we “shifted left” on cloud costs by treating attribution as a metadata and automation problem: using Backstage as the source of truth for service/team ownership, Terraform to enforce and propagate tags consistently, and Datadog Cloud Cost Management to turn that ownership into near-real-time, engineer-readable cost signals across AWS and key SaaS/data platforms.

You’ll see the patterns that made this scale: automated tag coverage and exception handling, normalization of cost dimensions into service/team/environment, and Infrastructure-as-Code generation of Datadog dashboards and monitors so every team gets a consistent cost cockpit by default. We’ll also cover what we wish we’d known at the start—why dashboards alone don’t drive action, how to design ownership-aware anomaly alerts that route to the right Slack channels, and how to build a feedback loop that turns cloud cost management from a quarterly finance exercise into an everyday engineering practice powered by Datadog.