Datadog APM & Distributed Tracing Fundamentals Practice Questions

Datadog APM & Distributed Tracing Fundamentals Certification exam preparation

Datadog APM & Distributed Tracing Fundamentals practice questions in English: 180 questions with explanations. The mock exam has 90 questions and 120 minutes.

Datadog APM & Distributed Tracing Fundamentals free question 9 / 10 · Domain 3: Insight Discovery

What retention period and unit of capture does the official documentation give for diversity sampling?

  1. It scans all ingested spans and keeps uniformly chosen spans in proportion to traffic volume for 15 days
  2. It scans error spans and keeps at least one per host every hour at most for 7 days, replacing old ones
  3. It scans root spans in order and keeps at least one per combination, such as environment and service, every minute at most, for 15 days, replacing old ones
  4. It scans service entry spans and keeps at least one per combination, such as environment and service, every 15 minutes at most, for 30 days
Show the answer and explanation

Answer: D

Why this is correct
The Diversity sampling part says it scans through the service entry spans and retains for 30 days: at least one span (and the associated trace) for each combination of environment, service, operation, and resource every 15 minutes at most; high latency spans for the p75, p90, and p95 per combination; and a representative selection of spans that ensures error diversity. The page notes that this dataset is not uniformly sampled and is biased toward errors and high-latency traces.

Why the other options are wrong
"It scans root spans in order and keeps at least one per combination, such as environment and service, every minute at most, for 15 days, replacing old ones" … It scans service entry spans, not root spans, at 15-minute intervals at most and for 30 days. Root spans, one minute, and 15 days all fail to match.
"It scans all ingested spans and keeps uniformly chosen spans in proportion to traffic volume for 15 days" … It scans service entry spans and retains for 30 days. The page states the dataset is not uniformly sampled and does not represent all traffic proportionally.
"It scans error spans and keeps at least one per host every hour at most for 7 days, replacing old ones" … The scan covers service entry spans, the unit is the combination of environment, service, operation, and resource, the interval is 15 minutes at most, and the period is 30 days. Error spans, per host, one hour, and 7 days do not match.

How it differs from similar items
Diversity sampling is a biased, representative sample, while one percent flat sampling is a uniform sample based on the trace ID. The first helps you find an example from a low-traffic endpoint, and the second suits trend analysis.

Go to the Datadog APM & Distributed Tracing Fundamentals practice app (10 free questions; purchase unlocks all questions)

See the full practice set