Zero Instrumentation Observability Based on eBPF

What if you could monitor your entire infrastructure metrics, logs, and traces without touching your application code? This session explores how eBPF is reshaping observability for modern, containerized environments.

Comprehensive observability is no longer optional for engineering teams that need to detect and resolve issues fast. But the traditional path to getting there is costly: extensive instrumentation across every application drains time and resources before you've gained a single insight. This talk examines a leaner approach.

We explore how eBPF enables automatic telemetry collection, including metrics, logs, and traces, directly at the kernel level, eliminating the instrumentation burden entirely. You'll see how it monitors container activity in real time, capturing network requests and filesystem interactions without any code changes required.

The session closes with practical strategies for putting that telemetry data to work: streamlining troubleshooting and turning raw observability signals into faster, more confident incident resolution.

If you're an SRE, DevOps engineer, or platform team lead looking to build deeper visibility into your distributed systems with less overhead, this talk delivers a hands-on look at what eBPF-powered, zero-instrumentation observability looks like in practice.