Explore Shopify's journey in migrating from a custom distributed tracing pipeline to OpenTelemetry in this 19-minute conference talk. Learn about the motivations behind the migration, the advantages of leveraging OpenTelemetry, and gain insights into the concrete steps of the migration process. Discover how Shopify implemented distributed tracing, including context propagation, custom instrumentation, and service onboarding. Examine the architecture of their tracing pipeline and the unexpected benefits they encountered. Delve into cross-language support, the transition through OpenCensus, and the final move to OpenTelemetry. Gain valuable lessons from Shopify's experience with the OpenTelemetry Collector, Resource Labeler, and the implementation of W3C Trace Parent.
Migrating to OpenTelemetry From a Custom Distributed Tracing Pipeline
CNCF [Cloud Native Computing Foundation] via YouTube
Overview
Syllabus
Introduction
Before OpenTelemetry
What we need to implement distributed tracing
When we started
Context propagation
Free options
Custom instrumentation library
Onramp for services
Architecture
Unexpected Benefits
CrossLanguage Support
Open Census
Collector
Open Sensors Ruby
Spoiler Alert
Moving to OpenTelemetry
OpenTelemetry Collector
Resource Labeler
OpenTelemetry in Rust
OTL Exporter
The Agent
The Strategy
W3C Trace Parent
Lessons Learned
Outro
Taught by
CNCF [Cloud Native Computing Foundation]
