Building software reliability requires an understanding that goes beyond writing good code and implementing tests; it involves actively ensuring that the software will perform as expected under various conditions. Distributed tracing is a key technique for achieving this, as it provides insights into the actual execution of code, thus transforming uncertainty into predictability. The article emphasizes the importance of instrumentation in development, which allows for the real-time observation of software behavior in production environments without disrupting operations. It also highlights the significance of focusing on the four golden signals—latency, traffic, errors, and saturation—from the customer's perspective to enhance system reliability. Challenges associated with distributed tracing in modern complex systems, such as the need for manual context propagation and understanding diverse technology stacks, are acknowledged. Finally, the article discusses the broader concept of observability, urging developers to adopt open standards like OpenTelemetry for a more seamless integration of metrics, logs, and traces in order to build a comprehensive understanding of their systems' performance and reliability.