Pixie: eBPF Observability for Kubernetes
Shriira Press
Instant, code-free observability for Kubernetes — eBPF captures every request, metric, trace, and CPU profile inside the cluster, queryable with a few lines of Python-like script.
Welcome to Pixie: eBPF Observability for Kubernetes.
Pixie is an open-source observability platform that gives you deep visibility into Kubernetes applications without changing a single line of code. It uses eBPF to automatically capture full-body requests, service metrics, network traces, and continuous CPU profiles directly in the kernel, stores that telemetry on each node inside the cluster rather than shipping petabytes off to a vendor, and lets you query all of it with PxL, a Pythonic scripting language. This book builds Pixie up one layer at a time. The opening chapters explain what Pixie is and how eBPF auto-instrumentation works, then how protocols are traced into structured data tables. From there we tour the architecture — the per-node PEM agents, the cluster-level Vizier, the Kelvin aggregator, and Pixie Cloud — and the in-cluster edge data store with its distributed query engine. Later chapters cover the PxL language, the Live UI and CLI, continuous CPU profiling and dynamic logging, and finally how teams adopt Pixie in practice. Originally built by New Relic and contributed to the CNCF in 2021, Pixie has become a fast, low-overhead foundation for understanding what your cluster is really doing.
This title is part of the ShriIra library and is free to read in full, right here — our small contribution to making world-class knowledge easy to reach.
A note on reading it: open the Contents menu at the top of the reader to jump between chapters, use the Aa menu to set a comfortable text size, theme (light, sepia, or night), and single- or two-page layout. Your place is saved automatically, so you can always pick up where you left off.
We hope it serves you well.
— Shriira Press