Why Tracing is a Lifesaver for E2E Tests of Distributed Systems?
summary
Claim: distributed tracing dramatically cuts root‑cause time for flaky end‑to‑end tests by showing the actual inter‑service transactions. The demo uses tundra foresight to surface logs, screenshots and a single trace map that highlights a red span between lambda and dynamodb, and it uses time travel debugging to reveal a failed save request. The result: engineers find the failing span without full repro, fix the bug faster, and measure test bottlenecks.
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:23Why do endtoend tests fail
- 2:39Debugging errors
- 4:44Tundra Foresight
- 5:38Trace Map
- 6:10Time Travel Debugging
- 6:30Summary
- 7:03Outro
related talks
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
49mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
22mDistributed Tracing
16mDistributed Tracing is Awesome, Testing is Not... Until Now
34mReduce Complexity with Distributing Tracing
42mDistributed Tracing Opening the Door to Visualisation and Correlation
14mLeveraging Distributed Tracing in Kubernetes & Containerized Environments
32mA Picture is Worth 1,000 Traces