eBPF in Production Observability: Maps, Probes, and the Cost of Seeing Everything
bpftrace, BCC, and production agents: kprobes versus tracepoints, map memory, and why CO-RE beat copying kernel headers into the cluster.
2 min read
Tag
4 articles · 8 min of reading
bpftrace, BCC, and production agents: kprobes versus tracepoints, map memory, and why CO-RE beat copying kernel headers into the cluster.
Why long-lived services bloat RSS, how arenas and size classes work, and when switching allocators is a real lever versus a blog meme.
Submission and completion rings, SQPOLL, and when io_uring beats epoll plus worker threads — plus the security history you should patch.
cycles, instructions, IPC, cache misses, and stalled-cycles: using Linux perf before you rewrite a method the profiler barely saw.