What is OpenTelemetry
OpenTelemetry (OTel) — CNCF-graduated vendor-neutral standard and tooling for collecting + exporting observability data. Merges OpenTracing + OpenCensus (2019). Supports: metrics, logs, traces via a single SDK for 15+ languages. Auto-instrumentation for popular frameworks (Express, FastAPI, Django, Spring). Export to any backend: Jaeger, Prometheus, Datadog, New Relic, Grafana Tempo.
Below: details, example, related terms, FAQ.
Details
- SDK (in app): generates spans, metrics
- Collector: receives from SDK, processes (sampling, filtering), exports
- Exporters: Jaeger, Prometheus, OTLP (open protocol), vendor-specific
- Auto-instrumentation: drop-in for Express, Flask, Django, FastAPI, Spring, Rails
- Context propagation: trace_id via HTTP headers (W3C Trace Context)
Example
// Node.js Express auto-instrumentation
const { NodeSDK } = require('@opentelemetry/sdk-node');
const { getNodeAutoInstrumentations } = require('@opentelemetry/auto-instrumentations-node');
const sdk = new NodeSDK({
instrumentations: [getNodeAutoInstrumentations()],
});
sdk.start(); // All Express routes auto-tracedRelated Terms
Learn more
Frequently Asked Questions
OpenTelemetry vs Datadog agent?
OTel — open standard, vendor-neutral. Datadog agent — vendor-specific, deeper integrations. OTel migration easier if you switch vendors.
Performance overhead?
2-5% CPU overhead with auto-instrumentation. Sampling (1-10% traces) reduces it.
Is it stable?
Tracing — stable since 2021. Metrics — stable since 2023. Logs — stable since 2024. All three usable in prod.