Prometheus as exposition format for eBPF programs ..
summary
An operator compiles and loads eBPF ELF programs into a Kubernetes cluster, reads BPF maps to extract counters, and exposes those values as Prometheus metrics consumed by Grafana dashboards. The demo counts packets by protocol across a three‑node cluster and shows per‑node rates, proving the pipeline works with negligible overhead but requiring recent kernels and a GPL license in the ELF. The implementation uses clang to build eBPF, an operator to deploy, and Prometheus exposition for scraping.
AI-written summary · about
chapters (15)
- 0:00Introduction and setup
- 1:49Speaker introduction
- 4:27What is eBPF and its origins
- 7:44How eBPF works in the kernel
- 15:03Advantages and limitations of eBPF
- 18:59eBPF example: counting packets
- 20:31Deploying eBPF in Kubernetes
- 23:26Building a Kubernetes eBPF operator
- 27:29Operator implementation details
- 28:56Demo setup in Kubernetes
- 30:21Live demo of packet counting
- 31:57Analyzing the collected metrics
- 32:38Related tools and projects
- 33:44Key takeaways
- 34:29Q&A session
related talks
18mFind out what your processes are capable of with ebpf
33mAuto-instrumentation of Prometheus for RED Monitorin...
27mStep By Step Kubernetes Observability with eBPF
29mKubernetes Observability with eBPF
42mMonitoring Kubernetes Traffic with eBPF
36mBumbleBee: Or How I Learned To Stop Worrying And Love eBPF
35mAn Introduction to eBPF - Superpowers for Networking, Observability & Security
01h 02mTutorial: Create Your First EBPF Program to Monitor Your System Using Bumblebee