OTel SDK and Instrumentation

Properties

created 14.04.2026, 17:00

modified 26.07.2026, 13:52

published Empty

topics OpenTelemetry, Observability

authors Empty

ai-assisted Yes

Language-specific libraries that instrument application code and emit telemetry ( OTel Language SDKs)
Available for Python, Go, Java, JavaScript/Node, .NET, Ruby, Rust, PHP, and more
The SDK is what runs inside your application process; the OTel Collectorruns outside

# Instrumentation Types

Zero code changes — hooks into frameworks and libraries automatically
Instruments HTTP servers/clients, DB drivers, message queues, gRPC, etc.
Language-specific agents or monkey-patching
- Python: opentelemetry-instrument python app.py
- Java: -javaagent:opentelemetry-javaagent.jar
- Node.js: --require @opentelemetry/auto-instrumentations-node/register
Best for: getting started quickly, brownfield apps

Explicitly create spans and add attributes in code

1
2
3
with tracer.start_as_current_span("my-operation") as span:
    span.set_attribute("user.id", user_id)
    result = do_work()

Pre-built plugins for popular libraries (requests, sqlalchemy, flask, express, etc.)
Import and register alongside the SDK

TracerProvider — factory for creating Tracer instances
MeterProvider — factory for creating Meter instances (metrics)
LoggerProvider — factory for creating Logger instances
Propagators — handle context injection/extraction across process boundaries (see Context Propagation)
Exporters — send data to Collector or directly to backends via OTLP