Effortless, standardised homelab observability with eBPF
summary
Deploying eBPF-powered Beyla plus OpenTelemetry and Prometheus generates unified HTTP, SQL and trace telemetry across languages without changing app code, but requires running as root, kernel >5.8 with BTF (had to recompile Raspberry Pi kernel, ~23 minutes) and a Helm two‑command install. It surfaces end‑to‑end traces and latency hotspots immediately, misses in‑process custom context and has semantic‑version gaps, so add synthetic probes via blackbox_exporter for dead‑node alerts; overhead is low in practice (~1–2 cores only under extreme load).
AI-written summary · about
chapters (11)
- 0:00Home lab observability intro
- 4:19Traditional monitoring limits
- 8:17Criteria for home lab solution
- 10:23OpenTelemetry and eBPF options
- 12:51eBPF instrumentation with BAI
- 15:41Deploying BAI in home lab
- 17:49SLOs for alerting pitfalls
- 19:59Synthetic monitoring for reliability
- 21:04Book catalog with observability
- 26:28Manual vs eBPF trade-offs
- 27:33Q&A on eBPF overhead
related talks
33mEffortless, Standardised Homelab Observability With eBPF
37meBPF: Revolutionizing Observability and Telemetry for Modern Applications
31mZero instrumentation observability with eBPF
35mAn Exemplary Path: Leveraging EBPFs and OpenTelemetry To Auto-instrument for Exemplars
19mDiscover seamless observability with eBPF
05mBeyla: eBPF automatic instrumentation
31meBPF for Observability: The Good, the Bad, and the Ugly
30mAgentless instrumentation with eBPF, Prometheus and OpenTelemetry