Efficient and Large-scale Infrastructure Monitoring With Tracing
summary
Demonstrates LTTng features that extract high‑precision metrics remotely and capture pre‑crash history: LTTng snapshots/flight recorder records in‑memory ring buffers and dumps on demand, producing e.g. 8,000 events in 2 seconds on a laptop; live streaming to a relay supports real‑time monitoring and live trace streaming to relay. Benchmarks show 257M events (6.8 GB) recorded with system‑wide tracing versus 61M (5.4 GB) for strace, and no measurable throughput loss when streaming to a remote relay or writing traces to a dedicated disk. Includes a top‑like viewer (ltng‑top) and kernel+user tracing with dynamic enable/disable and crash snapshot integration.
AI-written summary · about
chapters (5)
- 0:00Start
- 4:26Tracing session example
- 12:14Flight recorder session + snapshot
- 25:38Advanced KVM analysis
- 41:24Questions?
related talks
38mLTTng: Beyond Ring-Buffer Based Tracing
37mKernel and userspace tracing with LTTng and friends
36mLTTNG-UST Efficient System Wide User-Space Tracing
12mStriaTrace: Efficient Tracing and Diagnosis for Online LLM Inference
27mBeyond OTLP: Unlocking the Potential of OS-native Tracing
26mLTTng for full stack tracing
52mTracking Down What Happened: A Comparison of Tracing Solutions
35mLinux Tracing updates