Skip to content

OpenTelemetry Certified Associate (OTCA) Fact Sheet

Exam Overview

Exam Code: OTCA Exam Name: OpenTelemetry Certified Associate Level: Associate Duration: 90 minutes Format: Multiple choice and multiple select, online proctored Questions: 60 Passing Score: 75% Cost: USD 250 (includes one free retake) Valid For: 2 years Delivery: Online proctored through PSI Prerequisites: None

Verify before booking. CNCF exam details, pricing, and curriculum versions change. Confirm on the official pages below.

📖 OTCA certification page - registration and curriculum 📖 Linux Foundation OTCA page - exam logistics and candidate handbook 📖 OpenTelemetry documentation - the primary study source 📖 CNCF curriculum repository - published exam domains

Why this exam is in this repo

Until now the repo had topics/observability.md, an observability service comparison, an LLM observability comparison, and a monitoring stack build, but not a single observability certification. OTCA closes that gap, and it does so vendor-neutrally: OpenTelemetry is the instrumentation standard that Datadog, Grafana, Splunk, Honeycomb, New Relic, and every hyperscaler now consume.

That makes it more durable than a vendor certification. Learning OTel is learning the layer underneath all of them.

Target Audience

  • SREs and platform engineers who own the observability pipeline
  • Developers instrumenting services
  • Anyone running the OpenTelemetry Collector in production
  • Kubernetes practitioners extending PCA or CKA knowledge into telemetry

Assumed background: comfortable with containers, distributed systems basics, and at least one programming language.

Exam Domains

Domain 1: The OpenTelemetry API and SDK (46%)

Nearly half the exam. Instrumentation is the core skill.

Key Concepts: - The API and SDK split, and why the API alone is a no-op - Signals: traces, metrics, logs, and their data models - Traces: spans, span context, span kinds, attributes, events, links, status - Context propagation: the W3C Trace Context standard, traceparent and tracestate, and baggage - Propagators and how context crosses process boundaries - Metrics instruments: counter, up-down counter, histogram, gauge, and their synchronous and asynchronous variants - Aggregation, temporality (delta and cumulative), and views - Logs and the log data model, including correlating logs with traces - Resources and resource detection, and semantic conventions - Samplers: always on, always off, trace ID ratio, parent-based, and head versus tail sampling - SDK configuration: providers, processors, exporters, and environment variable configuration - Automatic (zero-code) instrumentation versus manual instrumentation - Instrumentation libraries and the OpenTelemetry Protocol (OTLP)

📖 OpenTelemetry concepts - signals, context, and the data model 📖 Language SDKs - per-language instrumentation

Domain 2: The OpenTelemetry Collector (26%)

Key Concepts: - Collector architecture: receivers, processors, exporters, connectors, extensions - Pipelines per signal type and how components compose - Deployment patterns: agent (per host or sidecar) and gateway (standalone cluster) - Common receivers: OTLP, Prometheus, Jaeger, Zipkin, filelog, hostmetrics, kubeletstats - Common processors: batch, memory limiter, attributes, resource, filter, transform, tail sampling, k8sattributes - Common exporters: OTLP, OTLP HTTP, Prometheus remote write, debug, and vendor exporters - Connectors that join pipelines, such as spanmetrics - Collector distributions: core, contrib, and building a custom distribution with the OpenTelemetry Collector Builder - Configuration structure and validation - The OpenTelemetry Operator for Kubernetes, including auto-instrumentation injection

📖 Collector documentation - architecture and configuration 📖 Collector components registry - available receivers, processors, exporters

Domain 3: Fundamentals of Observability (18%)

Key Concepts: - Observability versus monitoring, and why the distinction matters - The three signals and what each is good and bad at - Telemetry correlation: exemplars, trace-to-log and trace-to-metric linking - Cardinality: what drives it, why it is expensive, and how to control it - Semantic conventions and why standardized attribute names matter - SLIs, SLOs, and error budgets - The golden signals: latency, traffic, errors, saturation - RED and USE methods - Sampling strategy and its effect on what you can answer

📖 Semantic conventions - standardized attribute names 📖 Observability primer - concepts and vocabulary

Domain 4: Maintaining and Debugging Observability Pipelines (10%)

Key Concepts: - Diagnosing missing telemetry: instrumentation, propagation, collector, or backend - Collector internal telemetry and health checks - Debug and file exporters for local troubleshooting - Memory limiter behavior and backpressure - Queue and retry configuration, and what happens when a backend is unavailable - Broken trace context and where propagation typically fails - Cost control: sampling, filtering, attribute reduction, and metric cardinality limits - Versioning, stability guarantees, and migration from legacy agents

Signal quick reference

Signal Answers Cost driver Watch out for
Traces Why is this request slow, and what did it touch Span volume Sampling decisions change what you can debug later
Metrics Is the system healthy, is it trending Time series cardinality One unbounded attribute can multiply series enormously
Logs What exactly happened in this code path Volume and retention Unstructured logs are hard to correlate