Keynote: Distributed Tracing - The Struggle is Real
summary
Distributed tracing often fails to deliver because adoption falls off after day one, leaving a few experts to wrangle complex tools like Jaeger while most engineers ignore them. Four recurring problems are clear: power-user tooling only, heavy sampling erodes trust, and instrumentation has no ownership, plus tracing is frequently tacked onto existing workflows instead of integrated. Practical use is to tie traces into alerts and dashboards, use traces to narrow the investigation haystack, and feed trace data into SLOs and incident workflows so it actually reduces mean time to resolution.
AI-written summary · about
chapters (6)
- 0:00Distributed tracing struggles
- 2:34Tracing often fails in practice
- 3:58Adoption problems in tracing tools
- 6:07Beyond technical features to purpose
- 7:51Tracing as complexity reducer
- 9:58Driving adoption through integration
related talks
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
51mNavigating Distributed Traces
28mTracing the Journey into Distributed Tracing
34mReduce Complexity with Distributing Tracing
49mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
38mDistributed Tracing in Large Software Systems
45mMastering Distributed Tracing: Insights from Building a Platform from Scratch
14mLeveraging Distributed Tracing in Kubernetes & Containerized Environments