Retroactive Sampling with OpenTelemetry: Cut 90% Distributed Traci... Roman Khavronenko & Zhu Jiekun
About this talk
This talk discusses the challenges of traffic consumption associated with tail-sampling traces in large-scale applications that operate across multiple clusters and availability zones. The speakers, Roman Khavronenko and Zhu Jiekun from VictoriaMetrics, introduce a retroactive sampling approach that enhances bandwidth efficiency by allowing OpenTelemetry collectors to buffer spans and send only the necessary data to a centralized trace backend. By implementing this strategy, attendees will learn how retroactive sampling can significantly reduce costs associated with tail-sampling and gain insights into its architectural design, practical implementation, and trade-off analysis for existing systems.
More from this event
See all 436 talks →
Best of KubeCon + CloudNativeCon Amsterdam 2026
2:17
The Quiet Work of Forever: Sustaining Open Source Communities - O. Hope Amaechi-Okorie, JSON Schema
26:24
Evolving KServe: The Unified Model Inference Platform for Both Predictive and... F. Spolti & J. Lee
32:40
Preventing S3 Cost Storms: Applying Cortex’s Efficiency Lessons to I/O-Heav... A. Fishman-Lichterman
5:32