Zero Instrumentation Observability with eBPF

What if you could collect metrics, logs, and traces across your entire infrastructure without adding instrumentation to a single application? This session explores how eBPF is making comprehensive observability faster and less resource-intensive for modern engineering teams.

Building reliable observability has always come with a significant time and resource cost. Adding instrumentation to each application individually creates friction that slows teams down and makes full-stack visibility difficult to maintain at scale.

This talk walks through how eBPF enables telemetry collection at the kernel level, covering how to monitor container activity including network requests and filesystem operations, without the traditional instrumentation overhead.

The session then moves into practical territory, sharing concrete strategies for using metrics, logs, and traces to troubleshoot and resolve problems effectively across distributed systems and containerized infrastructure.

If you're a DevOps engineer or SRE looking to build deeper visibility into your Kubernetes environment while reducing the operational burden of observability, this talk offers a practical, experience-driven look at what eBPF makes possible.