Panel: The Spec-tacular Game Show V2
summary
A fast, correction-driven panel that quizzes on concrete OpenTelemetry rules: API vs SDK roles, instrumentation scope not library, async metric instruments cannot attach exemplars while sync ones can, the simple span processor exports synchronously (avoid in production), span context is immutable, default metric cardinality is 2,000, OTLP uses no compression by default, resource detectors may return errors, and auto-instrumentation works without adding the SDK as a dependency.
AI-written summary · about
chapters (19)
- 0:00Welcome to the game show
- 1:24Game show rules
- 3:31OpenTelemetry API vs SDK
- 5:43Metric instruments and exemplars
- 7:34Span processors behavior
- 9:01Span context immutability
- 10:07Metric cardinality limits
- 11:50Server attribute conventions
- 13:15OpenTelemetry Protocol defaults
- 14:58OpenTelemetry profiles compatibility
- 16:23Resource detectors and errors
- 17:49Real or Halulu API challenge
- 21:26Auto instrumentation needs
- 22:28Partial identifying rules
- 23:11Global error handler
- 24:58OTLP signal and baggage
- 26:06Context propagation
- 27:30OpenTracing compatibility
- 29:21Closing and pavilion
related talks
50mPanel: Observability
46mPanel: Meet The Maintainers
34mPanel: OpenTelemetry: Bridging Platform and Enablement
01h 00mPractical OpenTelemetry in Javascript/Typescript
34mPANEL: OpenTelemetry Technical Committee Panel
30mPanel: Telemetry That Matters
36mPanel Discussion: The Present and Future of Open Source Observability
24mObservability and instrumentation via OpenTelemetry