Ingesting everything isn't a strategy, it's a bill that keeps growing, dashboards nobody trusts, and a compliance posture that's one audit away from a very bad week. Bryan Pierson from Enterprise Observability Engineering at US Bank shares how his team moved from ingesting everything-just-in-case to running a deliberate, multi-destination observability pipeline that actually reflects what the business cares about. He walks through how Datadog Observability Pipelines became the foundation for something more than log forwarding but a platform for filtering noise before it hits ingest, sampling high-volume low-value sources, scrubbing sensitive data at the edge, and routing logs to the right destination for the right reason. He also covers the piece most teams skip: accountability. A custom quota program tied directly to the pipeline gives every application team a log budget based on resiliency tier with automated enforcement that scales from a friendly incident at 80% all the way to disabled log flow at 120%. The result? Teams that care about what they log. Costs that behave. And an observability practice that stopped worrying and started working.

 42      0