X-Rays
FinOps X Logo

Engineering a Cost per Feature Capability with Open Telemetry

Agenda / Engineering a Cost per Feature Capability with Open Telemetry
Breakout
Level: 300
Unit Economics
Unit Economics

In this session, we share the journey of delivering “Cost per Feature” information that bridges the gap between raw cloud spend and product value. We will detail a methodology that combines cloud billing data with distributed tracing to attribute costs at the feature level—even across complex, asynchronous backend service chains.

Key takeaways for attendees:
The Methodology: How to use Open Telemetry trace span duration as a proxy for compute resource consumption and precise token counts for AI service attribution.
Connecting the Stack: Strategies for linking frontend user interactions to backend infrastructure costs using a unified Telemetry SDK and trace context propagation.
Navigating Accuracy & Sampling: Managing the trade-offs between data volume and precision, including how to statistically adjust for trace sampling to provide reliable directional insights.
Driving Business Value: Practical applications for feature-level data, from evaluating the ROI of premium features and informing pricing models to incentivizing engineering efficiency.

Speakers