The text discusses the challenges of using distributed tracing for long-term monitoring due to its high cost and diminishing returns over time. It highlights the benefits of using metrics instead, which provide a more cost-effective way to monitor system health and performance. The article also explores how transforming trace data into metrics can help solve problems such as capacity planning, alerting, and identifying performance bottlenecks. It introduces the OpenTelemetry Collector, a tool that simplifies the process of converting traces into metrics using an extract, transform, and load pipeline. The article explains the differences between the span metrics processor and connector, with the latter being a more recent and improved solution. It also provides guidance on configuring the span metrics connector, including implementing sampling policies to control cost. Overall, the text aims to demonstrate how converting traces into metrics can enhance observability and monitoring strategies, leading to significant cost savings and improved system performance.