I85PREP
⌘ K
OTCA study materials
OTCA study resources

OTCA practice questions

One question at a time. Build your OTCA understanding.

First 50 of 198 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1OpenTelemetry Fundamentals and Concepts

Which OpenTelemetry signal records discrete events that occurred at a specific point in time, often with an associated severity level?

  1. Profiles
  2. Logs Correct answer.
  3. Metrics
  4. Traces

Explanation

Logs are timestamped records of discrete events, and OpenTelemetry's log data model allows them to carry severity, body, and attributes. Metrics aggregate measurements over time, and traces follow a request across components.

Question 2OpenTelemetry Fundamentals and Concepts

In OpenTelemetry, what is a span?

  1. A collection of related traces grouped by service name
  2. A metric instrument that records cumulative totals
  3. A single operation within a trace, with a start time, end time, and metadata Correct answer.
  4. A configuration file that defines exporters

Explanation

A span represents one unit of work in a trace and captures timing plus contextual attributes. Spans are linked by parent-child relationships to form a trace.

Question 3OpenTelemetry Fundamentals and Concepts

Which component is responsible for collecting, processing, and exporting telemetry in the OpenTelemetry architecture?

  1. The Collector Correct answer.
  2. The context propagator
  3. The API
  4. The instrumentation library

Explanation

The OpenTelemetry Collector is a standalone service that receives, processes, and exports telemetry data. It can also be deployed as an agent or gateway between applications and backends.

Question 4OpenTelemetry Fundamentals and Concepts

What does the OpenTelemetry API provide to application developers?

  1. Interfaces and no-op implementations for instrumenting code Correct answer.
  2. A dashboard for visualizing telemetry
  3. A full agent that automatically instruments all libraries
  4. A storage backend for traces and metrics

Explanation

The API supplies the interfaces used to create telemetry and includes no-op implementations so instrumentation is safe even when no SDK is configured. The SDK implements those interfaces.

Question 5OpenTelemetry Fundamentals and Concepts

Which propagation mechanism does OpenTelemetry define as its standard for carrying trace context in HTTP headers?

  1. W3C Trace Context Correct answer.
  2. AWS X-Ray trace header
  3. Jaeger uber-trace-id
  4. B3 single header

Explanation

OpenTelemetry specifies W3C Trace Context as its default propagation format, using the traceparent and tracestate headers. Other formats are supported through optional propagators.

Question 6OpenTelemetry Fundamentals and Concepts

What is the role of a resource in OpenTelemetry?

  1. It defines sampling rules for traces
  2. It describes the entity producing telemetry, such as a service or host Correct answer.
  3. It routes telemetry between Collector pipelines
  4. It stores spans until they are exported

Explanation

A resource is a set of attributes describing the source of telemetry, such as service.name and host.name. These attributes are attached to telemetry to identify its origin.

Question 7OpenTelemetry Fundamentals and Concepts

Which metric instrument should be used to record a value that can go up and down, such as the number of items in a queue?

  1. Gauge observation
  2. Counter
  3. UpDownCounter Correct answer.
  4. Histogram

Explanation

An UpDownCounter records a value that can increase or decrease, making it suitable for quantities like queue depth. A Counter only increases.

Question 8OpenTelemetry Fundamentals and Concepts

What is the purpose of baggage in OpenTelemetry?

  1. To sample traces at a fixed rate
  2. To compress span data before export
  3. To propagate key-value pairs across service boundaries independent of traces Correct answer.
  4. To encrypt telemetry in transit

Explanation

Baggage carries application-defined key-value pairs through the context so downstream services can read them. It is separate from trace context and is not automatically added to spans.

Question 9OpenTelemetry Fundamentals and Concepts

An application creates a span but no SDK is registered. What happens to the span?

  1. It raises an error that must be handled by the caller
  2. It is automatically exported to a default endpoint
  3. It is created as a non-recording span and produces no exported data Correct answer.
  4. It is buffered in memory until an SDK is registered

Explanation

Without an SDK, the API returns a no-op span that records nothing and exports nothing, so instrumentation remains safe. This allows libraries to be instrumented without requiring an SDK at runtime.

Question 10OpenTelemetry Fundamentals and Concepts

What distinguishes a span's parent-child relationship from a span link?

  1. A link creates a new trace, while parent-child merges traces
  2. Parent-child is set by the Collector, while links are set by the API
  3. Parent-child expresses causal nesting within a trace, while a link associates spans across traces Correct answer.
  4. Parent-child is used only for metrics, while links are used for traces

Explanation

Parent-child relationships build the tree structure of a single trace and imply causality. Links associate a span with other spans, often in different traces, without implying nesting.

Question 11OpenTelemetry Fundamentals and Concepts

Which span kind should be used for a span that represents handling an incoming remote request?

  1. Producer
  2. Server Correct answer.
  3. Client
  4. Internal

Explanation

The Server kind marks a span that processes an inbound remote call, such as an HTTP handler. Client spans represent outbound calls, and Internal spans represent in-process operations.

Question 12OpenTelemetry Fundamentals and Concepts

What is the primary function of a span processor in the OpenTelemetry SDK?

  1. To sample decisions at the trace root
  2. To receive ended spans and forward them to an exporter Correct answer.
  3. To parse incoming HTTP headers
  4. To define metric aggregation temporality

Explanation

A span processor intercepts spans when they start and end and passes ended spans to exporters. Batch and simple processors are common implementations.

Question 13OpenTelemetry Fundamentals and Concepts

Which OpenTelemetry Collector component is responsible for receiving data?

  1. Exporter
  2. Connector
  3. Receiver Correct answer.
  4. Processor

Explanation

Receivers accept data into the Collector in a specific format or protocol. Processors transform data, exporters send it onward, and connectors join pipelines.

Question 14OpenTelemetry Fundamentals and Concepts

A team wants to drop spans with a specific attribute before export using the Collector. Which component type should they configure?

  1. An exporter in the traces pipeline
  2. A processor in the traces pipeline Correct answer.
  3. A receiver in the traces pipeline
  4. A service extension

Explanation

Processors sit between receivers and exporters and can filter or transform telemetry. Dropping spans based on attributes is a processor responsibility, such as the filter processor.

Question 15OpenTelemetry Fundamentals and Concepts

Which statement best describes head-based sampling?

  1. Sampling is applied only to metrics, not traces
  2. Each span is sampled independently by the exporter
  3. The sampling decision is made when the trace begins and is propagated to downstream services Correct answer.
  4. The decision is made after the full trace is assembled at the backend

Explanation

Head-based sampling decides at the root of the trace and communicates the decision through trace context so all services honor it. Tail-based sampling instead decides after seeing the complete trace.

Question 16OpenTelemetry Fundamentals and Concepts

What does the OpenTelemetry semantic conventions project provide?

  1. A reference implementation of a tracing backend
  2. A programming language for writing exporters
  3. Standardized attribute names and values for common telemetry Correct answer.
  4. A wire protocol for encoding spans

Explanation

Semantic conventions define agreed attribute names such as http.request.method so telemetry is consistent across systems. They are not a backend, protocol, or language.

Question 17OpenTelemetry Fundamentals and Concepts

Which protocol is defined by OpenTelemetry as its native telemetry transport?

  1. Prometheus remote write
  2. OTLP Correct answer.
  3. StatsD
  4. Syslog

Explanation

OTLP is the OpenTelemetry Protocol, designed natively for transmitting traces, metrics, and logs. It supports gRPC and HTTP encodings.

Question 18OpenTelemetry Fundamentals and Concepts

Which instrument is appropriate for recording request durations so that percentiles can be computed?

  1. Histogram Correct answer.
  2. Counter
  3. Observable gauge
  4. UpDownCounter

Explanation

A Histogram records the distribution of values in buckets, enabling percentile estimation. Counters and up-down counters track sums, not distributions.

Question 19OpenTelemetry Fundamentals and Concepts

Why does OpenTelemetry recommend using the API in library code rather than the SDK directly?

  1. It automatically enables tail-based sampling
  2. It keeps the library decoupled so applications can choose their own SDK and configuration Correct answer.
  3. It removes the need for any exporter
  4. It guarantees lower CPU usage than the SDK

Explanation

Depending only on the API lets a library emit telemetry without forcing a specific SDK or configuration on its users. The application decides which SDK to install and how to export.

Question 20OpenTelemetry Fundamentals and Concepts

A service receives a request with a traceparent header and must continue the same trace. What must it do?

  1. Export the header directly to the backend without creating spans
  2. Create a new root span and link it to the header value
  3. Start a new trace and ignore the incoming header
  4. Extract the remote context and create a new span as a child of the extracted span context Correct answer.

Explanation

The service should extract the remote span context from the header and use it as the parent for its new spans. This preserves trace continuity across service boundaries.

Question 21OpenTelemetry Fundamentals and Concepts

Which statement about logs in OpenTelemetry is accurate?

  1. Log records can carry a trace ID and span ID to correlate them with traces Correct answer.
  2. Logs cannot include structured attributes
  3. Logs must be exported using a proprietary protocol
  4. Logs bypass the SDK and are written directly to disk

Explanation

The log data model allows log records to reference a trace ID and span ID, enabling correlation between logs and traces. Logs also support structured attributes and flow through the SDK.

Question 22OpenTelemetry Fundamentals and Concepts

What is the difference between a synchronous and an asynchronous metric instrument?

  1. Synchronous instruments run in the Collector, while asynchronous ones run in the application
  2. Synchronous instruments are only for traces, while asynchronous ones are only for logs
  3. Synchronous instruments record measurements at call time, while asynchronous ones report values via callbacks during collection Correct answer.
  4. Synchronous instruments require OTLP, while asynchronous ones require Prometheus

Explanation

Synchronous instruments like Counter and Histogram are updated directly when events occur. Asynchronous instruments use callbacks to observe values at collection time, such as observable gauges.

Question 23OpenTelemetry Fundamentals and Concepts

An organization needs to sample traces based on errors that only become visible after all spans of a trace are collected. Which approach fits?

  1. Tail-based sampling in the Collector Correct answer.
  2. Always-on sampling with no sampler
  3. Head-based sampling in the SDK
  4. Random sampling at the exporter

Explanation

Tail-based sampling buffers complete traces and decides afterward, allowing policies based on errors or latency. Head-based decisions occur before the outcome is known.

Question 24OpenTelemetry Fundamentals and Concepts

How does the Collector's connector component differ from an exporter?

  1. A connector only handles metrics, while an exporter only handles traces
  2. A connector receives external traffic, while an exporter parses headers
  3. A connector writes to disk, while an exporter writes to memory
  4. A connector passes data between pipelines inside the Collector, while an exporter sends data to an external destination Correct answer.

Explanation

Connectors join two pipelines, acting as an exporter for one and a receiver for another within the same Collector. Exporters instead transmit data to systems outside the Collector.

Question 25OpenTelemetry Fundamentals and Concepts

Why is it important that trace context propagation preserves the trace flags field?

  1. It sets the metric aggregation temporality
  2. It identifies the Collector instance
  3. It carries the sampling decision so downstream services record spans consistently Correct answer.
  4. It encrypts the trace payload

Explanation

The trace flags in the traceparent header include the sampled bit, which tells downstream services whether to record. Preserving it keeps sampling consistent across the trace.

Question 26OpenTelemetry Fundamentals and Concepts

Which three signal types does OpenTelemetry define as its core observability signals?

  1. Traces, profiles, and events
  2. Metrics, alerts, and dashboards
  3. Traces, metrics, and logs Correct answer.
  4. Logs, traces, and configuration

Explanation

OpenTelemetry's specification centers on three signals: distributed traces, metrics, and logs, each with its own API and SDK components. Profiles are a separate, emerging effort rather than one of the three established core signals.

Question 27OpenTelemetry Fundamentals and Concepts

What is the fundamental unit of work recorded by OpenTelemetry's tracing API?

  1. A metric data point
  2. A span Correct answer.
  3. A resource
  4. A log record

Explanation

A span represents a single operation within a trace, capturing its name, timing, attributes, and relationships to other spans. Traces are composed of spans arranged in parent-child or linked structures.

Question 28OpenTelemetry Fundamentals and Concepts

A developer needs to emit telemetry from a shared library that will be used by many applications. Which dependency should the library take on?

  1. A vendor-specific tracing agent
  2. The OpenTelemetry API package Correct answer.
  3. The OpenTelemetry SDK for a specific language
  4. The OpenTelemetry Collector

Explanation

Libraries should depend only on the OpenTelemetry API, which is a no-op when no SDK is installed and delegates to the configured SDK when one is present. This keeps libraries vendor-neutral and avoids forcing a particular SDK or backend on the applications that use them.

Question 29OpenTelemetry Fundamentals and Concepts

Which statement describes how context propagation works across a process boundary in OpenTelemetry?

  1. The SDK writes span data to a shared database that downstream services query
  2. The receiving service generates a new trace ID and links it to the caller's trace after the fact
  3. The trace ID and span ID are injected into the outgoing request's carrier and extracted by the receiving service Correct answer.
  4. The Collector forwards the full span tree to each downstream service before the request is sent

Explanation

Propagation relies on inject and extract operations: the sending process serializes trace context into a carrier such as HTTP headers, and the receiver deserializes it to continue the same trace. This works without any shared storage or coordination through the Collector.

Question 30OpenTelemetry Fundamentals and Concepts

A team wants to reduce trace volume while guaranteeing that every trace containing an error is retained. Which sampling architecture meets this requirement?

  1. Head-based sampling with a higher probability on services that log errors
  2. Per-span sampling that drops only spans whose status is not error
  3. Head-based sampling with a fixed probability applied at the root service
  4. Tail-based sampling performed after all spans of a trace have been observed Correct answer.

Explanation

Tail-based sampling buffers spans until a trace is complete, so the sampling decision can consider the whole trace, including error status discovered late. Head-based decisions are made at trace start, before error information exists.

Question 31Signals: Traces, Metrics, Logs, and Baggage

Which OpenTelemetry signal is best suited to answering the question 'How many requests per second is the checkout service handling right now?'

  1. Traces
  2. Metrics Correct answer.
  3. Baggage
  4. Logs

Explanation

Metrics are aggregated numeric measurements designed for cheap, efficient querying of rates, counts, and distributions over time. Traces record individual request paths, logs record discrete events, and baggage carries cross-cutting key-value context.

Question 32Signals: Traces, Metrics, Logs, and Baggage

Which field links a log record to the trace it was emitted during?

  1. The trace ID and span ID carried on the log record Correct answer.
  2. The log body text
  3. The log timestamp
  4. The service name resource attribute

Explanation

Correlating logs with traces relies on the trace ID and span ID being attached to the log record so backends can join them. The body, resource attributes, and timestamp do not identify a specific trace.

Question 33Signals: Traces, Metrics, Logs, and Baggage

What is baggage in OpenTelemetry?

  1. A log sampling strategy
  2. A type of metric aggregation
  3. A buffer of unsent spans on the exporter
  4. Key-value context propagated across service boundaries alongside a trace Correct answer.

Explanation

Baggage is a standardized set of key-value pairs that travels with the request context so downstream services can read values without them being part of the trace itself. It is not a storage buffer, metric aggregation, or sampling policy.

Question 34Signals: Traces, Metrics, Logs, and Baggage

Which metric instrument type should be used to record a value that only ever increases, such as the total number of bytes sent?

  1. Gauge
  2. UpDownCounter
  3. Histogram
  4. Counter Correct answer.

Explanation

A counter is monotonic, meaning recorded values are non-negative and the total only grows. UpDownCounter allows decreases, histograms capture distributions, and gauges record instantaneous values.

Question 35Signals: Traces, Metrics, Logs, and Baggage

Which log severity level conventionally indicates the most severe condition?

  1. ERROR
  2. DEBUG
  3. WARN
  4. FATAL Correct answer.

Explanation

Severity levels are ordered, and FATAL represents a condition from which the application cannot continue, ranking above ERROR, WARN, and DEBUG. DEBUG is the least severe and typically used for development detail.

Question 36Signals: Traces, Metrics, Logs, and Baggage

What does the parent span ID field on a span represent?

  1. The next span to be created
  2. The span that directly caused or contains this span Correct answer.
  3. The root span of the entire trace
  4. The trace's sampling priority

Explanation

The parent span ID points to the immediate caller or enclosing operation, which is what gives a trace its tree structure. The root span is only the parent of its direct children, not of every span.

Question 37Signals: Traces, Metrics, Logs, and Baggage

Which best describes the relationship between a trace and its spans?

  1. A trace is a list of metrics sharing one name
  2. A trace is a tree of spans sharing one trace ID Correct answer.
  3. A trace is a single span with many logs
  4. A trace is a set of baggage keys

Explanation

All spans in a trace share the same trace ID and are connected by parent-child links, forming a tree rooted at the entry span. Metrics, logs, and baggage are distinct concepts.

Question 38Signals: Traces, Metrics, Logs, and Baggage

A service records request latency with a histogram instrument. What does this primarily enable?

  1. Guaranteeing exactly-once delivery of spans
  2. Computing percentiles and bucket-based distributions Correct answer.
  3. Propagating user IDs to downstream services
  4. Reducing log volume at the source

Explanation

Histograms aggregate observations into configurable buckets, from which quantiles and distributions can be derived in the backend. They do not affect span delivery, context propagation, or log volume.

Question 39Signals: Traces, Metrics, Logs, and Baggage

You want a metric that can rise and fall, such as the number of items currently in a queue. Which instrument fits best?

  1. Counter
  2. Exponential histogram
  3. Histogram
  4. Asynchronous UpDownCounter Correct answer.

Explanation

An UpDownCounter supports both increments and decrements, and the asynchronous variant observes a current value via callback, which suits queue depth. A plain counter cannot decrease, while histograms and exponential histograms record distributions.

Question 40Signals: Traces, Metrics, Logs, and Baggage

A developer adds a new attribute to a span after the span has already ended. What is the expected outcome?

  1. The attribute is converted into a log record
  2. The attribute is merged into the parent span
  3. The attribute is not recorded because the span is immutable once ended Correct answer.
  4. The span is automatically restarted

Explanation

Spans become immutable after End is called; the SDK will not accept further modifications. Attributes must be set while the span is active, and nothing is rerouted to the parent or converted into logs.

Question 41Signals: Traces, Metrics, Logs, and Baggage

Which propagation mechanism is defined by the W3C standard specifically for trace context?

  1. The authorization header
  2. The traceparent header Correct answer.
  3. The baggage header
  4. The content-type header

Explanation

W3C Trace Context defines the traceparent header carrying trace ID, span ID, and flags, with tracestate for vendor data. Baggage is defined in a separate W3C specification, and the other headers serve unrelated purposes.

Question 42Signals: Traces, Metrics, Logs, and Baggage

A team wants to pass a tenant identifier to all downstream services for routing, without making it part of the trace or metrics. Which mechanism is appropriate?

  1. Exemplars
  2. Resource attributes
  3. Baggage Correct answer.
  4. Span events

Explanation

Baggage is designed to propagate arbitrary key-value context across service boundaries independent of the trace structure. Span events and exemplars are trace-related constructs, and resource attributes describe the emitting entity rather than per-request context.

Question 43Signals: Traces, Metrics, Logs, and Baggage

What is the main risk of placing large or sensitive values in baggage?

  1. It is stored permanently in the collector's database
  2. It is propagated to every downstream service and may be exposed or bloat requests Correct answer.
  3. It converts traces into metrics
  4. It disables sampling for the trace

Explanation

Baggage travels with the request context to all downstream services, so secrets can leak and large payloads can inflate headers. It is not persisted by collectors, does not convert signals, and does not control sampling.

Question 44Signals: Traces, Metrics, Logs, and Baggage

A trace shows a child span with a parent span ID that does not exist in the collected data. What is the most likely explanation?

  1. The child span has an invalid trace ID
  2. The trace has more than one root by design
  3. The child span was created before its parent
  4. The parent span was not sampled or not exported Correct answer.

Explanation

Sampling and export decisions can drop a parent while a child is still recorded, producing orphan spans that reference a missing parent. A child cannot be created before its parent, and a valid trace ID is shared across the trace.

Question 45Signals: Traces, Metrics, Logs, and Baggage

Which statement about the OpenTelemetry log data model is correct?

  1. Log records cannot have attributes
  2. Log records are only produced by the collector
  3. Log records must always have a trace ID
  4. A log record can carry a severity, a body, attributes, and trace context Correct answer.

Explanation

The log data model defines fields including timestamp, severity, body, attributes, and optional trace and span IDs for correlation. Attributes are allowed, trace IDs are optional, and applications as well as collectors can emit records.

Question 46Signals: Traces, Metrics, Logs, and Baggage

An engineer wants to attach a discrete occurrence, such as an exception, to a specific span. Which span feature is intended for this?

  1. Span events Correct answer.
  2. Span kind
  3. Span status
  4. Span links

Explanation

Span events represent timestamped occurrences within a span's lifetime, and exceptions are conventionally recorded as events with exception attributes. Links associate spans across traces, status conveys success or failure, and kind describes the span's role.

Question 47Signals: Traces, Metrics, Logs, and Baggage

What is the purpose of the tracestate header in W3C Trace Context?

  1. To carry vendor-specific trace data alongside traceparent Correct answer.
  2. To store the full trace of spans
  3. To authenticate the collector
  4. To compress log payloads

Explanation

tracestate allows multiple vendors to propagate their own state in an ordered list while traceparent carries the core identifiers. It does not store spans, authenticate, or compress logs.

Question 48Signals: Traces, Metrics, Logs, and Baggage

A histogram metric is configured with buckets in milliseconds. An engineer queries the 99th percentile and finds it inaccurate near the top end. What is the most likely cause?

  1. The observation falls outside the defined bucket boundaries Correct answer.
  2. The exemplar filter removed samples
  3. The trace was not sampled
  4. The counter was reset

Explanation

Percentile accuracy from histograms depends on bucket boundaries; values above the highest boundary are clamped into an overflow bucket, losing precision. Counter resets, exemplar filters, and trace sampling do not affect histogram bucket resolution.

Question 49Signals: Traces, Metrics, Logs, and Baggage

Why might a team prefer an exponential histogram over an explicit bucket histogram?

  1. It removes the need for any aggregation
  2. It provides dynamic, high-resolution bucketing with lower configuration effort Correct answer.
  3. It automatically creates spans
  4. It can only record integer values

Explanation

Exponential histograms scale bucket resolution with the magnitude of values, giving good relative accuracy without hand-tuning boundaries. They still require aggregation and do not create spans or restrict values to integers.

Question 50Signals: Traces, Metrics, Logs, and Baggage

A service records a metric with a high-cardinality attribute such as a unique user ID. What is the primary consequence?

  1. The histogram becomes a gauge
  2. The attribute is silently dropped by the SDK
  3. The metric backend may suffer from series explosion and increased cost Correct answer.
  4. The metric becomes a trace automatically

Explanation

Each distinct attribute combination creates a new time series, so high-cardinality attributes can explode series counts and storage costs. The SDK does not convert metrics to traces or change instrument types, and attributes are not automatically dropped.