The Spec-tacular Game Show
summary
A fast paced game show that reads wrong OpenTelemetry claims and fixes them on the spot: expect clarifications like spec limits memory consumption (no unbounded buffers), OTLP default is HTTP proto not gRPC, and log body is optional (logs can lack a body); you also get concrete corrections on signals (more than traces/metrics/logs, baggage is an extension signal), attribute rules (keys non-empty), resource immutability, and that SDKs must not throw unhandled exceptions after boot.
AI-written summary · about
chapters (18)
- 0:00Welcome to the game show
- 1:04Game rules and contestants
- 2:07Memory limits in spec
- 3:13Signals beyond traces metrics logs
- 5:00Traces versus spans
- 6:02Logs and optional body field
- 7:51Metrics instruments complexity
- 9:40Attribute key rules
- 12:13Semantic conventions naming
- 14:00OTLP partial success HTTP codes
- 15:26OTLP transport protocols
- 16:50OpAMP protocol types
- 18:17Views and environment variables in metrics
- 19:46Game recap feedback
- 20:28Practice prompt OTLP exporter
- 21:33Resources and immutability
- 23:06Why resources immutable
- 25:36Global error handling SDKs
related talks
32mAntipatterns in Observability: Lessons Learned and How OpenTelemetry Solves Them
30mOpenTelemetry 101
12mSpec-Tacular Streaming: Lessons Learned in Standards Land
26mSpecTec, a Single Source of Truth
09mOpenTelemetry or OpenTeleMessy?
36mObservability For You and Me with OpenTelemetry
22mBridging the OpenTelemetry Adoption Gap in Cloud-Native Applications
15mOpenTelemetry: The Vision, the Reality and How To Get Started