Debugging Your Services with Distributed Tracing
summary
Concrete account of rolling out distributed tracing at Heroku: proven white‑box tracing using Zipkin on Heroku, instrumentation via rack middleware + Faraday to propagate trace/parent IDs and timestamp annotations, transport done out‑of‑band with a threaded client (sucker_punch), practical sampling (start near 0.1 sample rate), run time was a few months part‑time, Zipkin uses a test Cassandra and needs auth via reverse proxy, language support and host‑agent requirements determine whether the solution is viable.
AI-written summary · about
chapters (39)
- 0:00Start
- 2:16A Simple Use Case
- 3:24Microservices!
- 4:01Why do we need distributed tracing?
- 4:27Internal services look like external APIs
- 5:25How do you tell the story?
- 5:54What's Stopping You?
- 6:08Outside the Ruby Wheelhouse
- 6:42Fractured Ecosystem
- 7:04Theory - Practice
- 7:17The Basics
- 7:56Why might this not work for you?
- 9:07White Box Tracing
- 9:20Metadata Propagation
- 9:43History Lesson
- 9:56Dapper
- 11:58Tracing Requests
- 14:10A Trace is many Parent - Child Relationships
- 16:09Transporting the Data
- 16:55Viewing the Data
- 21:48Tracing Outgoing Requests
- 22:18More Middleware!
- 24:09Buy, Build, or Adopt
- 24:55Adopt an OSS Solution?
- 25:38What about Open Tracing?
- 26:05Standardizes Instrumentation
- 26:41Where is Open Tracing at today?
- 27:17Interoperability is Still Messy
- 27:42Language Support
- 28:42What are other folks doing?
- 29:06Infra Requirements and Limitations
- 30:02Installing a Separate Agent
- 30:36Missing Authentication & Authorization
- 30:47Client Authorization
- 31:07Browser Authentication
- 31:25Giving people access
- 31:41Custom Instrumentation
- 32:06What happens when data leaks?
- 32:46Get it on the Roadmap
related talks
30mDistributed Tracing: From Theory to Practice
36mDistributed Tracing: From Theory to Practice
50mDistributed Tracing in the Wild
31mTroubleshoot Your RoR Microservices with Distributed Tracing
29mDistributed Request Tracing
42mAn introduction to distributed tracing and Zipkin
24mDistributed Tracing : New DevOps Foundation.
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
