Strategies for Auto-Instrumenting Go Applications Without Code Changes

Explore comprehensive strategies for implementing auto-instrumentation in Go applications without requiring source code modifications in this 20-minute conference talk. Discover how to overcome Go's static nature and lack of dynamic code modification capabilities through ready-to-use, open-source solutions that enable instrumentation at both compile-time and runtime. Compare leading approaches including eBPF-based solutions, compile-time instrumentation techniques, and runtime methods such as shared library injection and binary trampoline implementations, while analyzing their trade-offs in performance, stability, and security. Learn how OpenTelemetry community collaboration has transformed auto-instrumentation from a theoretical concept into a practical reality for Go applications. Examine ongoing Go runtime developments that could unlock new auto-instrumentation patterns, including enhancements in function call interception, runtime hooks, and tracing APIs as outlined in golang/go#63185. Gain benchmark-driven insights to help evaluate the most suitable auto-instrumentation approach for production environments by balancing observability requirements, performance overhead, and deployment complexity. Access actionable insights and production-ready open-source tools that can be immediately integrated into Go applications, making this session valuable for SREs, performance engineers, and Go developers seeking to enhance their application monitoring capabilities.