True Kubernetes Observability with eBPF
summary
Claims that eBPF-based agents provide true out‑of‑band Kubernetes observability: an eBPF agent runs in the kernel, captures spans and API calls without changing apps, performs variant‑depth capture and early distributed digestion to cut storage, and keeps an in‑cloud data plane so sensitive traces never leave the cluster. Presenter cites surprising numbers — 20–30% of infrastructure spend on observability and ~180 microservices per company — and describes Ground Cover as deployed and working with minute‑scale rollout.
AI-written summary · about
chapters (9)
- 0:00Kubernetes observability challenges
- 3:01Legacy observability failures
- 8:46Developer and SRE gap
- 13:02eBPF for app observability
- 19:26eBPF architecture benefits
- 23:01Centralized observability limits
- 26:54Cost-efficient data collection
- 28:18Privacy in cloud-native
- 30:27Modern observability vision
related talks
30mObservability in Kubernetes: Unlocking the Power of eBPF
27mStep By Step Kubernetes Observability with eBPF
24mMaking the Kernel Speak Kubernetes: Unlocking eBPF’s Power for Observability
42mMonitoring Kubernetes Traffic with eBPF
37mZero-instrumentation observability based on eBPF
19mDiscover seamless observability with eBPF
31mZero instrumentation observability with eBPF
31mBridging Dev and Ops with eBPF: Extending Observability Upwards and Downwards