~/talk

Application Telemetry Through the Kernel With Linux user_events

links:none on file
slides:none on file

Linux user_events lets instrumented apps register tracepoints and write structured logs into per-CPU kernel/perf buffers so a listener can drain CPU-local rings without app-side queues. Microsoft builds an end-to-end pipeline: a user_events exporter for the OTel Rust SDK and a user_events receiver for the Otel-Arrow engine, pinned readers drain the same CPU. It enables on-the-fly log collection without restart, preserves locality and backpressure, but rings are bounded, require covering all CPUs to avoid silent loss, and need shared schema/decoding.

AI-written summary · about

  1. 0:00Introduction and speaker background
  2. 1:25Problem with static telemetry costs
  3. 3:13Use case for dynamic log enabling
  4. 4:57What are Linux user events
  5. 6:43Three key ideas behind user events
  6. 9:50How user events work end-to-end
  7. 13:04Cost comparison with OTLP
  8. 14:27OpenTelemetry integration
  9. 16:12Differences from OTLP pipeline
  10. 17:57Kernel ring buffer limits
  11. 19:21Security and privacy considerations
  12. 20:22When to use user events
  13. 21:05Trade-offs and requirements
  14. 21:48Key takeaways and summary
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.