Schedule

This talk explores how eBPF and Rust can overcome limitations to enable low-overhead observability and distributed tracing in native systems—without modifying application code. We’ll walk through a Rust-based eBPF implementation that observes function calls, injects and propagates trace context (e.g., traceparent), and emits OpenTelemetry spans from the kernel to user space using ring buffers. We’ll also show how this enables auto-instrumentation of mixed-language applications, where context flows seamlessly across C++, Rust, and other runtimes. If you’re building high-performance systems, this session will show how to bring OpenTelemetry distributed tracing deep into the native stack.