If that sounds familiar, you’re not alone. At some point, nearly every engineering or platform team working with Datadog hits this moment: a sky-high Datadog bill that doesn’t match the value you’re seeing. These Datadog pricing issues often stem from unchecked telemetry practices.

And the worst part? It’s often nobody’s fault.

You’re collecting Datadog logs, custom metrics, and traces because it’s the right thing to do. You need visibility, accountability, and the ability to troubleshoot fast. But under the hood, there’s a silent force driving up costs, slowing down dashboards, and making your observability stack harder to manage: cardinality in Datadog.

In simple terms, metric cardinality refers to the number of unique tag combinations your telemetry generates. And when you tag everything — env, region, user_id, container_id, and more — your seemingly simple metric like api.request.count can explode into millions of unique time series.

Multiply that across Datadog metrics, logs, and Datadog APM traces, and you’ve got: