Lightning Talk: Distributed Tracing Without the Pain!
summary
Approach reconstructs end-to-end traces without app changes by combining Envoy-based service mesh observations (span logs, request parameters, headers) with timing analysis, topology and constraint solving to map incoming requests to backend calls. Prototype outputs most-likely mappings plus per-trace confidence scores, stores short-lived logs for offline matching, and handles high load better than ordering baselines (96% accuracy vs 77%). It targets cases where header propagation is absent but notes limits with heavy async, caching or batching patterns.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:30Debugging microservices is hard!
- 1:10End-to-end distributed tracing
- 2:43Main challenge of end-to-end tracing
- 3:38How end-to-end tracing happens today: Header Propagation
- 4:53What information can we get from Envoy without app instrumentation?
- 11:44Summary: distributed tracing without instrumenting app
related talks
12mLightning Talk: With Great Traces Comes Great Costs: How to Reduce That Bill?
05mLightning Talk: That Small Difference Made all the Difference- 12mLightning Talk: With Great Traces Comes Great Costs: How to Reduce That Bill?
22mDistributed Tracing
24mDistributed Tracing : New DevOps Foundation.
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
09mGraduated Project Lightning Talk: Using Trace Data for Monitoring & Alerting of Appl...
59mAn introduction to distributed tracing