Tracing the Kubernetes Control Plane with eBPF and OpenTelemetry
Watch the full recording: how eBPF-based tracing diagnoses control plane outages that metrics alone can't explain.
About This Session
When a Kubernetes control plane outage hits, most teams are left debugging with only metrics. Built-in distributed tracing exists in the control plane, but enabling it means manual configuration and restarting core components.
eBPF moves trace collection and exporting out of process entirely, producing full end-to-end traces without an in-process OTel SDK and without restarting anything. Mike Dame and Ron Federman walk through how it works, from eBPF mechanics to real-world scenarios like webhook failures and control plane outages that metrics alone can't explain.
SEE IT FOR YOURSELF
Full system visibility in action
Complete end-to-end traces, no restarts, no code changes.
