What is this OpenTelemetry thing?
summary
OpenTelemetry claims to unify tracing logs and metrics via a standard OpenTelemetry Protocol plus SDKs for 11 languages, semantic conventions that normalize attribute names, and a configurable OpenTelemetry Collector that centralizes routing, redaction, and dual‑send. Examples: Prometheus/InfluxDB for metrics, Zipkin/Jager and ELK/log stash histories, Honeycomb as a backend. The talk explains coded vs auto instrumentation, memory/cost tradeoffs, and concrete sampling techniques like head sampling and tail sampling with a 20M‑span free tier example.
AI-written summary · about
chapters (19)
- 0:00OpenTelemetry intro
- 1:05Observability origins
- 2:30Metrics era tools
- 4:37Cloud complexity & TSDBs
- 6:23Logs & ELK evolution
- 7:46Distributed tracing standards
- 9:16What is OpenTelemetry
- 12:30OpenTelemetry core signals
- 13:52Tracing for distributed systems
- 19:49Logs in observability
- 21:38Metrics & time-series
- 24:51Tracing propagation
- 26:54Baggage in propagation
- 30:26Auto-instrumentation
- 32:11Coded instrumentation
- 34:19OpenTelemetry collector
- 37:28Sampling strategies
- 41:20OpenTelemetry future
- 42:23Resources & offerings
related talks
47mWhat Is This OpenTelemetry Thing?
20mObservability and Open Telemetry
22mOpenTelemetry – An Observability Framework for Cloud-Native Software
32mOpenTelemetry feat. logs, metrics and traces
30mOpenTelemetry 101
46mThe State of OpenTelemetry
35mOpenTelemetry for Busy People
20mKeynote: (Open)Telemetry Makes Observability Simple