Distributed Tracing: How the Pros Debug Concurrent and Distributed Systems
summary
Shows how distributed tracing preserves trace context across services using OpenTracing and collects spans into Zipkin (or Jaeger) via HTTP or Kafka, storing traces in Cassandra for offline Spark queries. Demonstrates real traces (e.g., 8 spans, 44 ms; long traces up to 355 ms), explains b3 headers and envelope propagation for message architectures, covers scope managers, out‑of‑band reporting, sampling to limit memory cost, and practical span tagging and log correlation so you can pinpoint which service and span creates latency.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:14What We're Going to Cover
- 2:46Monoliths and Microservices
- 3:41Microservices are for Scaling People
- 5:49Microservice People Problems
- 10:18Enter Distributed Tracing
- 12:28Distributed Tracing Features
- 16:35Collecting Distributed Traces
- 18:52Tracing Output
- 21:27Popular Tracing Implementations
- 24:19Open Tracing Terminology
- 27:32Collecting Trace Data (Code)
- 39:15Trace Propagation
- 41:06Best Practices . Standardize on carrier formats inside your services
related talks
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
34mReduce Complexity with Distributing Tracing
22mDistributed Tracing
37mObservability, Distributed Tracing & the Complex World
38mDistributed Tracing: Practices, Patterns & Technologies for Debugging Microservices
42mDistributed Tracing Opening the Door to Visualisation and Correlation
24mDistributed Tracing : New DevOps Foundation.
42mAn introduction to distributed tracing and Zipkin