Logs are valuable sources of information but managing billions generated daily is impractical and expensive. By extracting metrics from logs using Observability Pipelines, teams can retain important information while minimizing costs associated with log management. This approach enables the reduction of high-volume logs without compromising analytics capabilities or developer efficiency. Metrics extracted from logs are stored in Datadog for 15 months, allowing for predictive forecasting and seasonality checks. Observability Pipelines can also reduce log size before sending them to a destination, decreasing network and egress costs while maintaining optimal resource performance.