Observability has matured significantly by 2025, with OpenTelemetry (OTel) becoming a staple in many organizations. While initial enthusiasm around OTel focused on its potential to simplify observability, companies are now grappling with practical challenges as they refine their use of telemetry. Key issues include managing the costs associated with telemetry data, ensuring the quality of that data, and effectively managing the OTel Collector. One major concern is the impact of zero-code instrumentation. While it makes it easy to add telemetry without altering application code, it often leads to unnecessary data collection, which can inflate observability costs. The author points out the importance of maintaining high-quality telemetry while avoiding excessive noise. OpenTelemetry does offer ways to limit unwanted data, but the process varies by programming language, creating complexity. The author expresses a desire for a more streamlined approach to manage this auto-instrumentation. Telemetry costs also escalate with the volume of traces collected. Although traces are essential for understanding application flow, sending every trace can be wasteful. Many organizations lack the knowledge to implement sampling effectively, which could help manage costs. The author suggests that pre-configured sampling profiles for the OTel Collector could simplify this process. Furthermore, there's a growing emphasis on instrumentation quality, highlighted by the launch of the Instrumentation Score by OllyGarden, which aims to standardize and improve telemetry practices across the industry.