OpenTelemetry
Vendor-neutral standard for traces, metrics, and logs across cloud services.
About
OpenTelemetry (OTel) is the CNCF-graduated standard for instrumenting applications with traces, metrics, and logs. Instead of locking into a vendor SDK, developers emit telemetry through OTel's APIs and SDKs and export it via OTLP to any backend: Datadog, Honeycomb, Grafana, or open-source collectors. It is the second-most-active CNCF project and the default instrumentation layer for modern distributed systems.
OTel is also becoming the interchange layer for AI observability. Its GenAI semantic conventions define a shared vocabulary for LLM spans (gen_ai.operation.name, token usage, tool calls, agent runs), though as of mid-2026 every gen_ai attribute still carries Development rather than Stable status, so attribute names keep shifting and versions must be pinned. The conventions moved to a dedicated open-telemetry/semantic-conventions-genai repository in 2026, and gen_ai.system was renamed to gen_ai.provider.name in v1.37.0.
What it can do
- Vendor-neutral traces, metrics, and logs
- Auto-instrumentation for popular frameworks
- OTLP export to any observability backend
- GenAI semantic conventions for LLM spans
- Collector for processing and routing telemetry
Setup
Add the OTel SDK for your language, configure auto-instrumentation, and point the OTLP exporter at your backend.
Prerequisites
An application to instrument; a telemetry backend or collector.
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 88; Trending score 86/100 from researcher assessment.