OpenTelemetry Certified Associate (OTCA)
Validate foundational and practical expertise in open-source distributed observability using OpenTelemetry, including the API vs SDK architecture, OpenTelemetry Collector pipelines, distributed traces, metrics, logs, and auto-instrumentation.
Official Exam Specification & Quick Facts
Question Difficulty Breakdown
Exam Guidelines & Simulation Rules
Frequently Asked Questions (FAQ)
The OTCA is a multiple-choice certification from the CNCF and Linux Foundation validating foundational competence in OpenTelemetry standards, instrumentation, collector pipelines, and telemetry data collection.
The OTCA is a 90-minute multiple-choice examination consisting of 60 questions administered online with remote proctoring.
A passing scaled score of 75% is required. Results are displayed immediately upon submission.
Domains include: (1) OpenTelemetry Overview & Architecture (22%), (2) Telemetry Signals: Traces, Metrics, Logs (26%), (3) OpenTelemetry Instrumentation (22%), and (4) OpenTelemetry Collector (30%).
The API provides the programming interface and abstractions (Tracer, Meter, Logger) for application developers to instrument code without specifying an implementation. The SDK implements the API, managing sampling, batching, and exporting to backends.
The Collector is a proxy that receives, processes, and exports telemetry. Its pipeline components are: Receivers (accept data), Processors (transform, batch, filter), Exporters (send to backends), and Connectors (bridge pipelines).
Semantic conventions define standardized naming rules and attribute keys for telemetry data (e.g., http.status_code, db.system, net.peer.name) across all programming languages and frameworks.
A Span represents a single unit of work (with a name, timestamps, attributes, events). A Trace is a directed acyclic graph (DAG) of Spans representing the complete end-to-end journey of a request across services.
Context propagation serializes trace identity (TraceID, SpanID, TraceFlags) into wire protocol headers (W3C Trace Context: traceparent and tracestate) transmitted across network boundaries.
Automatic instrumentation uses bytecode manipulation or monkey patching (e.g. Java agent, eBPF) to inject telemetry with zero code changes. Manual instrumentation uses the API directly to capture fine-grained business logic.
Agent Mode (co-located on the same host or pod sidecar gathering local telemetry) and Gateway Mode (centralized, load-balanced standalone cluster handling heavy processing, routing, and backend export).
SREs, software engineers, DevOps specialists, platform architects, and observability engineers seeking vendor-neutral telemetry pipelines.